An Investigation to Resolve the Interactions among SOFC, Power-Conditioning Systems, and Application Loads

Project Investigators Sudip K. Mazumder, Kaustuva Acharya, and Rajni K. Burra

(University of Illinois) Michael von Spakovsky, Doug Nelson, Diego Rancruel, and Saravanan Subramanian

(Virginia Tech) Comas Haynes and Robert Williams

(Georgia Tech.) Joseph Hartvigsen and S. Elangovan

(Ceramatec Inc.) Dan Herbison and Chuck McKintyre

(Synopsys Inc.)

SECA Core Technology Program Review MeetingSeptember 30 - October 1, 2003

Albany, New York

SOFC Power-Conditioning System Modeling

gPROMS

(Balance-of-Plant System)

FORTRAN

(SOFC)

(Finite-Element Analysis)

PES

DC-DC DC-ACSOFC

BOPS

LumpedVoltageSource DC-DC

SOFC

BOPS

LumpedHarmonic

Load

iSight(System Integration)

SABER DESIGNER

(Power Electronics System)

Lumped Harmonic

Load

Impact of PES on SOFC

Impact of SOFC+BOPS on PES

Multi-Software Simulation Platform

AL

Comprehensive System Model

Reduced-Order Models

Transient SOFC Response to Electrical Stimulus: Modeling Approach

Reactants inlet flow rates and properties are invariant during relatively short transient episode

Fuel stream effects are dominant

Quasi-steady-state electrochemistry

Lagrangian extension of validated steady-state model to track fuel parcels that travel over electroactive area

Fuel Cell

11

Fuel Stream

Oxidant Stream

x

t + ∆t t

x+ ∆x

t t + ∆t v1

v2

ηelement(t+∆t) =ηfield(x+∆x, t +∆t)

Potentiostatic Control (Power Increase)

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)

Attainment of steady state at the time constant T = Lcell/vfuel

Impact of Electrical Stimulus:Galvanostatic Control (Power Increase)

Duality of potential drop seen: polarization curve effect and subsequent fuel depletion effect

Multiple voltage reductions are “seen” by the reactant streams

Transient is thus longer by multiples of the time constant

Larger initial fuel utilizations prolong the relative transient due to enhanced fuel depletion effects

0.60

0.625

0.65

0.675

0.70

0.725

0.000 0.500 1.000 1.500 2.000 2.500 3.000 3.500 4.000 4.500

τ(s)

Fuel Util.= 50%

Fuel Util.= 75%

Fuel Util.= 62.5%

Cel

l Vol

tage

(V)

A 20% rise in the load current

0.000 0.500 1.000 1.500 2.000 2.500 3.000 3.500 4.000 4.500

τ

Fuel Util.= 50%

Fuel Util.= 75%

Fuel Util.= 62.5%

A 20% rise in the load current

0.60

0.625

0.65

0.675

0.70

0.725

0.000 0.500 1.000 1.500 2.000 2.500 3.000 3.500 4.000 4.500

τ(s)

Fuel Util.= 50%

Fuel Util.= 75%

Fuel Util.= 62.5%

Cel

l Vol

tage

(V)

A 20% rise in the load current

0.000 0.500 1.000 1.500 2.000 2.500 3.000 3.500 4.000 4.500

τ

Fuel Util.= 50%

Fuel Util.= 75%

Fuel Util.= 62.5%

A 20% rise in the load current

Leveraging Approach to Planar Cells

- Initial attempt at seed simulation of vertical team developmental cell GE, SECA Annual Mtg., 4/03, 4 3/8” diameter cell

- Initial and final conditions match reported data

- Trends corroborate those of the tubular results

Seed Case Study: Potentiostatic Stimulus (0.8V to 0.6V)

0100200300400500600700800

0 0.2 0.4 0.6 0.8 1

Dimensionless Time

Cur

rent

Den

sity

[m

A/c

m2]

010203040506070

Fuel

Util

izat

ion

[%]

Current Density Fuel Util.

RAM Issues of “Real World” Operating Conditions

2 2CO CO C s→ + ( )CO H H O C s+ → +2 2 ( )

Carbon Monoxide Axial Profiles

2.00E-05

2.40E-05

2.80E-05

3.20E-05

3.60E-05

4.00E-05

0.00 0.20 0.40 0.60 0.80 1.00 1.20

Dimensionless Axial Distance

CO

flow

reat

e (m

ol/s

)

0 to 11 to 11 to 21 to 3

Frozen CO electrochemistry promotes coking via heightened presence of CO along the anode.

electrolyte

cathode

anode

A CB

F

D

GE

Leverage of SECA Failure Analysis and Lab efforts into characterization of impact of electrical conditions upon cell electrochemistry

0 5 10 15 20 25 300

100

200

300

400

500

5%20%30%

Fracture Toughness 2(J/m )cG

(J sec)q

Maximum allowable heating rate for cracking vs. fracture toughness

Microcrack Nucleation

BOPS Modeling: Summary

!Development of dynamic heat transfer, thermodynamic, kinetics, and physical models for each component of the BOPS:

"Compressor, expander, heat exchangers, steam generator, reformer, and fuel storage

!Implementation of models in a dynamic-programming environment using state-of-the-art transient numerical solver

!Integration of BOPS component models into a BOPS sub-system model

!Parametric studies (trade-off analyses) of best-practice control strategies for continuous operation and start-up and shut-down

BOPS: PARAMETRIC STUDIES

Pre-reformer Thermal Transient Response for FU Perturbations (SMR = 3.4 & FRR = 0.30 constant)

5

10

15

20

25

1 to 2 1 to 3 1 to 4 1 to 5 5 to 4 5 to 3 5 to 2 5 to 1

Power Changes (KW)

Tim

e (s

ecs)

FU = 0.85 (Steady Temp.) FU = 0.70 (Steady Temp.) FU = 0.90 (Temp.)

Power-demand and system-level parameter perturbations

Total system efficiency and power-demand and system-level parameter perturbations

0.20

0.30

0.40

0.50

1 2 3 4 5

Power Demand (KW)

Syst

em T

herm

al

Effic

ienc

y

FU = 0.85 FU = 0.70 FU = 0.90

Power demand perturbation Power-demand and system-level parameter

perturbations Small changes in power demand with floating

fuel utilization Power demand perturbation with

temperature control Total system efficiency analysis Start-up and shut-down

Outline

Methane Compressor Thermal Transient Response For SMR Perturbations (FU=0.85, FRR=0.3 constant)

0

2

4

6

8

1 to 2 1 to 3 1 to 4 1 to 5 5 to 4 5 to 3 5 to 2 5 to 1

Power Changes

Tim

e (s

ecs)

Thermal SMR=3.4 Thermal SMR =3 Thermal SMR=3.8

Power-demand and system-level parameter perturbations

BOPS: PARAMETRIC STUDIES

System Thermal Efficiency for Pre-reformer Inlet Gas temperature Control

0.27

0.32

0.37

0.42

5 4 3 2 1

Power Changes (KW)

Syst

em T

hal

Ef

ficie

nc

T4 = 1120K T4 = 1150K T4 = 1200K T4 = 1300KT4 = 1400K T4 = 1500K T4 = 1550K T4 = 1570K

Total system efficiency and power-demand perturbation with temperature control

-100

0

100

200

300

400

0 10 20 30 40 50 60 70 80 90 100

Hea

t Flu

x (k

W/m

2)

Steam Generator Segment

Heat Flux at O Sec. Heat Flux at 5O Sec.Heat Flux at 10O Sec. Heat Flux at 20O Sec.Heat Flux at 40O Sec. Heat Flux at 120O Sec.Max Allowed Heat Flux

Steam generator heat flux

300

400

500

600

700

800

900

0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200

Tem

pera

ture

(K)

Time

Stop recirculation at 700 K Stop recirculation at 750 KStop recirculation at 800 K Stop recirculation at 900 KStop recirculation at 600 K

Steam generator start-up

Steam-Methane Reformer Dynamic Reasponse for Start-Up (FU=0.85, SMR=3.4, FRR=0.3)

200400600800

10001200

0 500 1000 1500 2000 2500 3000

Time (sec)

Tem

pera

tue(

K)

0.E+00

4.E-05

8.E-050 10 20 30 40 50

Time (Sec)

Ref

orm

ate

Fel

(K

mol

/s)

Reformate Exit Temperature Reformer Wall TemperatureReformate Gases Mass Flow

Steam methane reformer start-up

0.47

erm yu

PES Modeling and SOFC PCS System Interactions: Summary

! Development of nonlinear switching models of PES! Integration of SOFC, PES, and BOPS models to develop a

comprehensive SOFC PCS model! Development of reduced-order models for fast and convergent

simulations on a PC! Investigated the impact of PES low- and high-frequency current

ripples and the effects of load-transients on SOFC performance! Analyzed the impact of SOFC-output-voltage variations on the

dynamics and stability of PES using bifurcation analysis! Investigated effects of PES control and modulation strategies on

SOFC performance! Conducted a preliminary trade-off study to determine the optimal size

of a energy-storage device (comprising a battery and a pressurized hydrogen fuel tank) to cost-effectively improve PCS transient response

! Designed a (low-cost) novel zero-ripple, energy-efficient, reduced-voltage-stress, and direct energy-conversion PES

PES Topological Models

Self-Commutated Voltage-Source InverterLine-Commutated Current-Source Inverter

°°°°°

n

Load3Φ

DC/ACBoost

°

Filter

Vin

Electric Utility

Zig-ZagTransformer

Electric Utility

SCR InverterDC Link

Vin

n

LOAD3Φ

°°°°°°

Filter

High-Frequency Transformer-Isolated Topology

n

n

Load3Φ

AC/ACBoost HF Inverter

°°°°°°

Filter ElectricUtility

Vin

BATTERY

SOFC PCS Steady-State Interaction Analyses

Hydrogen utilization increases with the increase in load

Current ripple has effect on the hydrogen utilization athigh load conditions

Low-frequency ripples do not necessarily lead to increased fuel utilization unless their magnitude is high

For high loads, rise in the temperature observed at low frequencies (high temperatures can cause interaction between SOFC electrolyte and electrodes leading to formation of high resistivity material and high microcrack densities)

Effect of Ripple Factor and Current Amplitude

Effect of Frequency and Current Amplitude

Novel SOFC PES

AC/AC HF Inverter

Zero Ripple Boost

Control

Power supply

FEATURES: Direct power conversion Reduced device voltage stress High energy efficiency Minimize SOFC output-current ripple Filter size and weight 50% less than conventional filter

Cost effective

FuelCell

Zero

-Rip

ple

Boo

st C

onve

rter

Bat

tery

HF

Inve

rter

AC

-AC

Filte

r3-ΦLoad

To U

tility

Grid

PES

SOFC Output Current

Novel SOFC PES

SOFC Output Current

0.8

0.84

0.88

0.92

0.96

1

0 0.1 0.2 0.3 0.4 0.5

Load (x 10000) (W)

Effic

ienc

y (x

100

) (%

)

Boost

HF Inverter

AC-AC

Overall0.8

0.84

0.88

0.92

0.96

1

0 0.1 0.2 0.3 0.4 0.5

Load (x 10000) (W)

Effic

ienc

y (x

100

) (%

)

Boost

HF Inverter

AC-AC

Overall

3 Phase 1 Phase

0.8

0.84

0.88

0.92

0.96

1

0 0.1 0.2 0.3 0.4 0.5

Load (x 10000) (W)

Effic

ienc

y (x

100

) (%

)

Boost

HF Inverter

AC-AC

Overall0.8

0.84

0.88

0.92

0.96

1

0 0.1 0.2 0.3 0.4 0.5

Load (x 10000) (W)

Effic

ienc

y (x

100

) (%

)

Boost

HF Inverter

AC-AC

Overall

3 Phase 1 Phase

SOFC PCS Load-Transient Interaction Analyses

0.5

0.6

0.7

0.8

0.9

1.0

0.00 0.01 0.02 0.03 0.04 0.05 0.06

Time (sec)

Voltage

0.5

0.6

0.7

0.8

0.9

1.0

0.00 0.01 0.02 0.03 0.04 0.05 0.06

Time (sec)

Voltage

Vol

tage

(V)

Time (sec)

1073

1074

1075

1076

0.00 0.01 0.02 0.03 0.04 0.05 0.06

Time (sec)

Tem

pera

ture

(K)

125 A

100 A

75 A

50 A1073

1074

1075

1076

0.00 0.01 0.02 0.03 0.04 0.05 0.06

Time (sec)

Tem

pera

ture

(K)

125 A

100 A

75 A

50 A

1070

1080

1090

1100

1120

1130

1140

1150

1160

1170

0 60 120 180 240 300 360 420 480 540 600

Time (sec)

Tem

pera

ture

(K)

1110

125 A100 A

75 A50 A

1070

1080

1090

1100

1120

1130

1140

1150

1160

1170

0 60 120 180 240 300 360 420 480 540 600

Time (sec)

Tem

pera

ture

(K)

1110

125 A100 A

75 A50 A

Temperature

(Cell)

(Stack)

TemperatureTemperature

(Cell)

(Stack)

Load-Transient Mitigation Techniques

Inverter Modulation Strategies

z (mm) x (mm) z (mm) x (mm)

Hyd

roge

n M

olar

Flo

w R

ate

Cur

rent

Den

sity

(A/s

q.m

m)

Res

pons

e Ti

me

(sec

)

Battery Size (Ah)Hydrogen Flow Rate (moles/sec)

Battery vs. Pressurized Hydrogen Tank

Time (sec)

Cur

rent

(A)

! Pressurized Hydrogen Fuel Tank and Battery

# Instantaneous supply of additional energy requirement from SOFC stack

! Inverter Modulation Strategies# Space-vector modulation (SVM) vs sine-wave

PWM (SPWM) used for the inverter # Battery acts as a stiff voltage source, providing

additional energy requirements during transients# Slower boost converter voltage-controller

response to prevent immediate change in SOFC energy demands

Nonlinear Hybrid Controller for DC-DC Boost ConverterLo

ad R

esis

tanc

e (Ω

)

Inpu

t Vol

tage

(V)

Out

put V

olta

ge (V

)

Inpu

t Cur

rent

(A)

Time (sec) Time (sec)

Time (sec)Time (sec)

! FEATURES# Hybrid control concept

based on combining integral-variable-structure control (IVSC) scheme and multiple-sliding-surface control (MSSC)

# Excellent steady-state and transient responses even under parametric variations and under perturbations of SOFC stack voltage and load

# Controller eliminates the bus-voltage error with a reduced control effort

# Control scheme can reduce the impact of very high-frequency dynamics due to parasitics on an experimental closed-loop system

Phase-II Objectives

! Develop and enhance fully transient nonlinear and temporal models for a variety of PES and BOPS components and for planar SOFC stacks

! Experimental validations of interaction-analyses results

! Develop capabilities for analyzing long-term performance and durability of SOFC planar and tubular stacks due to their system interactions with the PESsand application loads and the BOPSs

! Develop cost-effective optimal PES designs and design guidelines for (i) mitigation of electrical feedbacks on SOFC stack and (ii) technology transfer to SECA industry team

! Develop transient PES and PCS models and load profiles for vehicular APUs for performance and reliability analyses

! Develop optimal control and modulation strategies for robust PES and BOPS

! Develop decomposition techniques for optimizing PCS with respect to cost, reliability, size (power density), and response time

Acknowledgement

Department of Energy

Don Collins and Randall Gemmen (NETL)

Moe Khaleel (PNNL)

Daniel A. Norrick, Brad K. Palmer, Charles Vesely, and Todd Romine

(CUMMINS)

Tony Campbell and Nguyen Minh (GENERAL ELECTRIC)

Sean M. Kelly, John G. Noetzel, and K. S. Rajashekara (DELPHI)

Shailesh D. Vora (SIEMENS WESTINGHOUSE)

John Stannard (FUEL CELL TECHNOLOGY)

Jeff Bond (ENGINEOUS SOFTWARE INC.)

Jonathan Felton (PSE, UK)