Characterization of SOFC Electrolytes for Improved...

1
Characterization of SOFC Electrolytes for Improved Mechanical Robustness Characterization of SOFC Electrolytes for Improved Mechanical Robustness Ryan Berke, Bodhayan Dev, Mark Walter, The Ohio State University Ryan Berke, Bodhayan Dev, Mark Walter, The Ohio State University Mick Day, Michael Jansen, Scott Swartz, NexTech Materials Ltd Mick Day, Michael Jansen, Scott Swartz, NexTech Materials Ltd Elastic properties are obtained using ASTM 1875-08, a sonic resonance technique. Rectangular bar specimens are suspended in a furnace, and excited over a range of frequencies. Certain frequencies exhibit resonance. The resonant modes are reproduced via FEA. The frequencies are used to calculate Temperature-dependent elastic properties GOAL Actuator 240 3% mol YSZ ASTM 1875-08, a sonic resonance technique. and excited over a range of frequencies. The resonant modes are reproduced via FEA. Temperature-dependent elastic properties for 3-YSZ and 6-ScSZ. Configuration of Optimize FlexCell TM geometry Actuator & Sensor 200 220 GPa) 6% mol ScSZ Photo of resonance First flexural mode (F E ) Configuration of bar specimens used in the Optimize FlexCell geometry to improve mechanical robustness without sacrificing electrochemical Electrolyte Hanging Specimen 180 tic Modulus, E (G apparatus. experiments. without sacrificing electrochemical efficiency. Material 0.035 0.04 f4= 1884 Hz, 140 160 Elast 3 2 2 0.9465 1 6.585 E mf L t E b t L = + First torsional mode (F G ) efficiency. 0.02 0.025 0.03 0.035 f1 = 633 Hz (f E ) f2= 944 Hz f3= 1752 Hz f4= 1884 Hz, (f G ) 0 100 200 300 400 500 600 700 800 120 Temperature, T ( o C) Where E = Young’s Modulus (GPa) b t L Typical amplitude versus frequency data obtained via 0 0.005 0.01 0.015 m= Mass of specimen (g) L, b, t = Geometry of specimen (mm) f = Fundamental frequency in flexure (Hz) data obtained via resonance. First 5 resonant modes obtained using ANSYS. The elastic properties are used as inputs for the structural and thermal models below. INTRODUCTION 300 800 1300 1800 2300 f E = Fundamental frequency in flexure (Hz) First 5 resonant modes obtained using ANSYS. structural and thermal models below. A small-scale repeating geometry models the support Each small-scale geometry is modeled under uniaxial “Effective” properties are computed relative to those of “Effective” properties are validated with 4-point INTRODUCTION 100 Total Equivalent Stiffness δ A small-scale repeating geometry models the support structure. Each small-scale geometry is modeled under uniaxial tension using FEA. “Effective” properties are computed relative to those of a uniformly thick solid: “Effective” properties are validated with 4-point bending of representative samples. The FlexCell TM is the latest generation electrolyte- supported planar SOFC from NexTech Materials Ltd. 60 70 80 90 iffness (%) All Points Points having shared thickness δ Tensile displacement Reduction in stiffness with increased active supported planar SOFC from NexTech Materials Ltd. Support 20 30 40 50 60 Percent Equivalent Sti displacement applied to the model. increased active area. Support Structure 0 10 20 30 40 50 60 70 80 90 100 0 10 20 Percent Active Area (%) Specimen Geometry %AA %E eq Predicted Modulus (GPa) Experimental Modulus (GPa) model. % Active Area: Structure (GPa) (GPa) Large Hexes 57 39 79.6 73.3 Small Hexes 39 52 106.4 108.6 78 small-scale geometries were selected by varying: Principal stress distribution SOFC Construction and Operation Small Circles 36 55 111.7 112.1 No Hexes 0 100 204.6 202.2 1. Width of each hex, w (mm) 2. Space between hexes, b (mm) 3. Radius of corner fillets, r (mm) distribution resulting from edge Magnification of membrane stress with increased SOFC Construction and Operation 3. Radius of corner fillets, r (mm) 4. Thickness of membrane, tm (mm) 5. Thickness of support mesh, ts (mm) edge displacement. with increased active area. The “effective” properties can be used for larger scale modeling. FlexCell TM Innovation: A honeycomb-type structure provides a thicker support mesh with thin active area. scale modeling. Large-scale geometries are modeled using “effective” Large-scale geometry is modeled with shell elements The relatively stiffer ribs appear to carry more stress, Based on initial findings, recommendations are made Large-scale geometries are modeled using “effective” mechanical properties in active regions. Large-scale geometry is modeled with shell elements under structural and thermal loading using FEA. The relatively stiffer ribs appear to carry more stress, but unlike the active area the stresses are unmagnified. Based on initial findings, recommendations are made for the development of even larger cells. Rounded fillet to Stress response to out-of-plane Wider vertical rib reduce stress Narrow frame Whole Electrolyte color-coded to show “effective” to out-of-plane pressure loading. Whole Electrolyte show “effective” material types. The ribs also reduce out-of-plane displacements. Honeycomb structure in a FlexCell TM Electrolyte Stress response Narrow side rib Rounded fillet to reduce thermal Thin regions are more electro-chemically efficient for Honeycomb structure in a FlexCell TM 1 st principal stress contour plot of ultra large FlexCell with widened rib Each large-scale geometry was selected by varying: 1. Stiffness applied to active regions, Eeq (GPa) Stress response to horizontal temperature Narrow side rib stress Thin regions are more electro-chemically efficient for improved performance. 1. Stiffness applied to active regions, Eeq (GPa) 2. Support rib and frame dimensions, (mm) 3. Radius of corner fillets, (mm) temperature gradient. Large cells which are mechanically robust have a higher power density.. An individual cell model is used to evaluate the Current-collecting, corrugated foams in the cell transfer Foam properties in the model are changed over several Cell models consider variations in foam geometry to higher power density.. An individual cell model is used to evaluate the electrolyte within the context of a stack. Current-collecting, corrugated foams in the cell transfer loads onto the electrolyte. Foam properties in the model are changed over several iterations until the load-displacement behavior matches the experimental data. Cell models consider variations in foam geometry to characterize stress distributions in electrolytes. Mechanical properties of an Brittle electrolytes can withstand considerable bending Load The load-displacement behavior of foam compressed in properties of an ideal foam. Thick regions support against mechanical damage Brittle electrolytes can withstand considerable bending Transfer The load-displacement behavior of foam compressed in a load frame is reproduced in a contact model. Thick regions support against mechanical damage during manufacturing, assembly, and operation. Changes in foam geometry may reduce loads Load frame data. Contact simulation. Schematic of a planar SOFC stack assembly. Material property inputs. Load-displacement results. Changes in foam geometry may reduce loads (and therefore stresses) imposed on electrolyte.

Transcript of Characterization of SOFC Electrolytes for Improved...

Page 1: Characterization of SOFC Electrolytes for Improved ...netl.doe.gov/sites/default/files/event-proceedings/2011/seca/poster...Honeycomb structure in a FlexCell TM Stress response Narrow

Characterization of SOFC Electrolytes for Improved Mechanical RobustnessCharacterization of SOFC Electrolytes for Improved Mechanical RobustnessRyan Berke, Bodhayan Dev, Mark Walter, The Ohio State UniversityRyan Berke, Bodhayan Dev, Mark Walter, The Ohio State University

Mick Day, Michael Jansen, Scott Swartz, NexTech Materials LtdMick Day, Michael Jansen, Scott Swartz, NexTech Materials Ltd

Elastic properties are obtained using

ASTM 1875-08, a sonic resonance technique.

Rectangular bar specimens are suspended in a furnace,

and excited over a range of frequencies.

Certain frequencies exhibit resonance.

The resonant modes are reproduced via FEA.

The frequencies are used to calculate

Temperature-dependent elastic properties

GOALActuator

240

3% mol YSZ

ASTM 1875-08, a sonic resonance technique. and excited over a range of frequencies. The resonant modes are reproduced via FEA. Temperature-dependent elastic properties

for 3-YSZ and 6-ScSZ.

Configuration of

Optimize FlexCellTM geometry

GOALActuator & Sensor

200

220

Ela

stic

Mod

ulus

, E

(G

Pa)

6% mol ScSZ

Photo of

resonance

� First flexural

mode (FE)

Configuration of

bar specimens

used in the Optimize FlexCellTM geometry

to improve mechanical robustness

without sacrificing electrochemical

Electrolyte Hanging Specimen

180

Ela

stic

Mod

ulus

, E

(G

Pa)resonance

apparatus.

used in the

experiments.

without sacrificing electrochemical

efficiency.

Electrolyte

Material0.035

0.04

f4= 1884 Hz,

140

160

Ela

stic

Mod

ulus

, E

(G

Pa)

3 22

0.9465 1 6.585Emf L tE

b t L

= +

� First torsional

mode (FG)

efficiency.Material

0.02

0.025

0.03

0.035

f1 = 633 Hz

(fE)

f2= 944

Hz

f3= 1752

Hz

f4= 1884 Hz,

(fG)0 100 200 300 400 500 600 700 800

120

Temperature, T (oC)

Where

E = Young’s Modulus (GPa)

0.9465 1 6.585b t L

Typical amplitude

versus frequency

data obtained via

0

0.005

0.01

0.015

E = Young’s Modulus (GPa)

m= Mass of specimen (g)

L, b, t = Geometry of specimen (mm)

f = Fundamental frequency in flexure (Hz)

data obtained via

resonance.

First 5 resonant modes obtained using ANSYS.

The elastic properties are used as inputs for the

structural and thermal models below.

INTRODUCTION

300 800 1300 1800 2300fE = Fundamental frequency in flexure (Hz) First 5 resonant modes obtained using ANSYS. structural and thermal models below.

A small-scale repeating geometry models the support Each small-scale geometry is modeled under uniaxial “Effective” properties are computed relative to those of “Effective” properties are validated with 4-point INTRODUCTION100

Total Equivalent Stiffness δ

A small-scale repeating geometry models the support

structure.

Each small-scale geometry is modeled under uniaxial

tension using FEA.

“Effective” properties are computed relative to those of

a uniformly thick solid:

“Effective” properties are validated with 4-point

bending of representative samples.

The FlexCellTM is the latest generation electrolyte-

supported planar SOFC from NexTech Materials Ltd. 60

70

80

90

Per

cent

Equ

ival

ent

Stif

fnes

s (%

)

All Points

Points having shared thicknessδ

Tensile

displacement

Reduction in

stiffness with

increased active supported planar SOFC from NexTech Materials Ltd.

Support 20

30

40

50

60

Per

cent

Equ

ival

ent

Stif

fnes

s (%

)

displacement

applied to the

model.

increased active

area.

Support

Structure0 10 20 30 40 50 60 70 80 90 100

0

10

20

Percent Active Area (%)

Specimen Geometry %AA %Eeq

Predicted Modulus

(GPa)

Experimental Modulus

(GPa)

model.

% Active Area:

Structure (GPa) (GPa)

Large Hexes 57 39 79.6 73.3

Small Hexes 39 52 106.4 108.678 small-scale geometries were selected by varying: Principal stress

distribution

SOFC Construction and Operation

Small Hexes 39 52 106.4 108.6

Small Circles 36 55 111.7 112.1

No Hexes 0 100 204.6 202.2

1. Width of each hex, w (mm)

2. Space between hexes, b (mm)

3. Radius of corner fillets, r (mm)

distribution

resulting from

edge

Magnification of

membrane stress

with increased SOFC Construction and Operation 3. Radius of corner fillets, r (mm)

4. Thickness of membrane, tm (mm)

5. Thickness of support mesh, ts (mm)

edge

displacement.with increased

active area. The “effective” properties can be used for larger

scale modeling.

FlexCellTM Innovation: A honeycomb-type structure

provides a thicker support mesh with thin active area.

5. Thickness of support mesh, ts (mm) scale modeling.

Large-scale geometries are modeled using “effective” Large-scale geometry is modeled with shell elements The relatively stiffer ribs appear to carry more stress, Based on initial findings, recommendations are made provides a thicker support mesh with thin active area. Large-scale geometries are modeled using “effective”

mechanical properties in active regions.

Large-scale geometry is modeled with shell elements

under structural and thermal loading using FEA.

The relatively stiffer ribs appear to carry more stress,

but unlike the active area the stresses are unmagnified.

Based on initial findings, recommendations are made

for the development of even larger cells.

Rounded fillet to

Stress response

to out-of-plane

Wider vertical rib

Rounded fillet to

reduce stress

Narrow frame

WholeElectrolyte

color-coded to

show “effective”

to out-of-plane

pressure loading.

Whole

Electrolyte

show “effective”

material types.The ribs also reduce out-of-plane displacements.

Honeycomb structure in a FlexCellTM

Electrolyte

Stress response

The ribs also reduce out-of-plane displacements.

Narrow side rib

Rounded fillet to

reduce thermal

Thin regions are more electro-chemically efficient for

Honeycomb structure in a FlexCellTM

1st principal stress contour plot of ultra large FlexCell

with widened rib

Each large-scale geometry was selected by varying:

1. Stiffness applied to active regions, Eeq (GPa)

Stress response

to horizontal

temperature

Narrow side ribreduce thermal

stress

Thin regions are more electro-chemically efficient for

improved performance.

1. Stiffness applied to active regions, Eeq (GPa)

2. Support rib and frame dimensions, (mm)

3. Radius of corner fillets, (mm)

temperature

gradient. Large cells which are mechanically robust have a

higher power density..improved performance.

An individual cell model is used to evaluate the Current-collecting, corrugated foams in the cell transfer Foam properties in the model are changed over several Cell models consider variations in foam geometry to

3. Radius of corner fillets, (mm) higher power density..

An individual cell model is used to evaluate the

electrolyte within the context of a stack.

Current-collecting, corrugated foams in the cell transfer

loads onto the electrolyte.

Foam properties in the model are changed over several

iterations until the load-displacement behavior

matches the experimental data.

Cell models consider variations in foam geometry to

characterize stress distributions in electrolytes.

matches the experimental data.

Mechanical

properties of an

Brittle electrolytes can withstand considerable bendingLoad

The load-displacement behavior of foam compressed in

properties of an

ideal foam.

Thick regions support against mechanical damage

Brittle electrolytes can withstand considerable bendingLoad

TransferThe load-displacement behavior of foam compressed in

a load frame is reproduced in a contact model.

Thick regions support against mechanical damage

during manufacturing, assembly, and operation.

Transfer

Changes in foam geometry may reduce loads

Load frame data. Contact simulation.Schematic of a planar SOFC stack assembly. Material property inputs. Load-displacement results.

Changes in foam geometry may reduce loads

(and therefore stresses) imposed on electrolyte.