Manufacturing Model: Simulating Relationships Between Performance, Manufacturing, and Cost of Production
TIAX LLCAcorn ParkCambridge, Massachusetts02140-2390
Reference: TIAX LLC -80034DE-FC26-02NT41568
SECA Core TechnologyProgram Workshop
Sacramento
February 19-20, 2003
1
1 Technical Issues
2 R&D Objectives and Approach
3 Activities for Phase I
2
Technical Issues
For commercial success, SOFC technologies must ultimately be manufacturable and cost competitive. A number of factors contribute to uncertainty at this time.
Cell design, stack designs, and production processes are still in early stages of development
SOFC stacks are radically different in structure from any currently mass-produced ceramic productsRelationships between cell and stack design, design tolerances, and stack performance are not very well established
3
Technical Issues
Proposed manufacturing processes may be amenable to high-volume production, however, specific processes and sequences must be selected.
Tape Cast
AnodePowder Prep
VacuumPlasmaSpray
ElectrolyteSmall Powder
Prep
ScreenPrint
CathodeSmall Powder
Prep
Sinter in Air1400C Sinter in Air
ProgressiveRolling of
Interconnect
ShearInterconnect
VacuumPlasmaSpray
SlurrySpray
ScreenPrint
Slurry Spray
Slip Cast
Multi-FiredProcess
Finish Edges
Note: Alternative production processes appear in gray to thebottom of actual production processes assumed
BrazePaint Braze
ontoInterconnect
Blanking /Slicing
QC LeakCheck
Interconnect
Fabrication
Electrolyte CathodeAnode
Stack Assembly
Electrical layer powders are made by ball milling and calcining.Interconnects are made by metal forming techniques.Automated inspection of the electrical layers occurs after sintering.
Electrical layer powders are made by ball milling and calcining.Interconnects are made by metal forming techniques.Automated inspection of the electrical layers occurs after sintering.
Illustr
ative
Process Flow Assumptions
Process Flow Process Flow AssumptionsAssumptions
Multi-Fired Process FlowMultiMulti--Fired Process FlowFired Process Flow
Potential Process Flow for Planar Anode-Supported SOFC
4
Technical Issues
Relationships between cell and stack design, design tolerances, stack performance, and process yields are not very well established.
Properties of individual layers, e.g., physical attributes, conductivity (electrical or ionic), polarization, transport, mechanical, are not well defined as a function of temperature
Manufacturing OptionsIndividual process stepsSequence of steps
Impact onProcess yield, tolerances, and reproducibilityPerformanceThermal cycling and LifeCost
5
Technical Issues Challenges
A state-of-the-art SOFC manufacturing model will allow developers and NETL to minimize the uncertainties inherently associated with commercialization of a new technology. The model must be able to:
Handle all key SOFC stack components, including ceramic cells and interconnects
Relate manufactured cost to product quality and likely performance, taking into account
manufacturing tolerancesproduct yieldline speed
Address a range of manufacturing volumes, ranging from tens of MW to hundreds of MW per year
Adapt to individual production processes under development by SECA industrial teams
6
1 Technical Issues
2 R&D Objectives and Approach
3 Activities for Phase I
7
R&D Objectives and Approach Objectives
The Manufacturing Model Project will develop a tool to provide guidance to the DOE and SECA development teams on system design and manufacturing processes selection.
Phase ISOFC Manufacturing Model
Framework and Demonstration
Phase 2SOFC Manufacturing Model Expansion
and Use
Develop model framework
Demonstrate benefit of model for system development trade-off analyses
Develop Phase 2 plan
Expand Phase I model framework to other SOFC system designs, alternative materials, and manufacturing processes
Incorporate findings and research of SECA teams
Objecti ve s
Model framework
Demonstration of model capabilities
Workshop with SECA stakeholders
Deliverables
The primary output of the model will be an activity based manufacturing cost for various SOFC system scenarios.
8
Phase I will be conducted in three tasks.
Develop architecture of manufacturing model
Review architecture with SECA stakeholders
Revise existing model architecture based on Task 1 workshop
Demonstrate manufacturing model with baseline SOFC system
Report project progress
Prepare Phase I report that summarizes critical manufacturing steps and performance parameters
Define Phase II development effort
Workshop with SECA stakeholders
Definition of model framework, user interface with model, and critical issues to be assessed, model assumptions
Workshop with SECA stakeholders
Monthly updates
Phase I final report
Objecti ve s
Deliverables
Task 2Model
Demonstration
Task 3Reporting
Activities for Phase I Tasks
Task 1Model Framework
Development
9
Activities for Phase I Deliverables
We anticipate that we will provide DOE and industrial teams with some key conclusions and recommendations:
Identification of critical manufacturing steps and performance parametersif considerable uncertainty exists about these steps, specific additional SECA R&D objectives may be developed
Refinement of SECA technology cost and performance estimatesDefinition of desirable next steps
10
Model Architecture Modeling Approach Link to Performance/Structural Module
The cost model will be augmented with a SOFC performance model to help relate manufacturing quality to performance.
User Interface
SOFC ScenarioCompiler Module
Activity-Based CostModel
PerformanceStructural Module
Databases
11
Model Architecture Modeling Approach Cost Model
The model uses a set of databases to calculate cost for defined production (process flow) scenarios and performance assumptions.
ManufacturingCosts
MaterialPropertiesDatabase
MaterialCost
Database
PurchasedComponents
FormulationDatabase
ProcessDatabase
CapitalEquipmentDatabase
• Labor• Real estate• Overhead
.
.
.
• Vendors• Cost vs.
volume...
• Density• Particle size
distribution• Surface area
.
.
.
• Cost vs. volume
• Specifica-tions
.
.
.
• Anode• Cathode• Electrolyte• Interconnects
• Equipment process data
• Throughput• Size limit• Automation• Scrap• Yield
• Cost vs. product volume
• Process flow• Equipment options
Calculation Engine (Activity-Based)
• Design• Performance
Parameters• Manufacture
Processes and Flow• Production Scenarios
• Tables• Graphs• Crystal Ball
– sensitivity– “frequency
distribution”
Inputs Outputs (Results)
The model description provides a unified framework for discussion of input parameters of interest to the Team members.
12
Model Architecture Performance/Structural Module Capabilities
The module also accounts for all the relevant thermo-electrochemical phenomena which influence cell performance and, ultimately, cost.
Interconnect• Heat conduction• Current
conduction
Anode and cathode reaction zones• Electrochemical reactions• Heat generation
Electrolyte• Ion conduction• Heat conduction
Flow passages• Heat convection• Plug flow of gas
Anode and cathode porous electrodes• Heat conduction• Current conduction• Species diffusion • Internal reforming on anode
13
Model Architecture Performance/Structural Module Interface with Cost Model
The performance/structural module is used to predict power density, thermal stresses, and other performance factors that influence cost.
Compressive load on the cell#
Temperature gradients
Material yield
Heat conduction
Boundary conditions
Heat generation
Heat convection
Current generation
Chemical reactions*
Stack/cell geometry
Electrochemical reactions
Power density
Stress distribution Defects
Performance/Structural Module Framework
* Internal reforming reactions# Compressive load needed for establishing contact between different stack layers
14
1 Technical Issues
2 R&D Objectives and Approach
3 Activities for Phase I
15
Manufacturing Model Technical Issues
We met with the SECA technical teams to discuss what relationships among cell and stack design, design tolerances, stack performance, and process yields should be considered in Phase 1?
Properties of individual layersThickness and other physical attributesPolarization and conductivity (electrical or ionic)TransportMechanical
Manufacturing OptionsIndividual process stepsSequence of steps
Impact onProcess yield, tolerances, and reproducibilityPerformanceThermal cycling and lifeCost
16
Manufacturing Model Stack Design
We discussed selection of a stack design for demonstration of the model capabilities and an initial assessment of the impact of selectedmanufacturing/design factors.
What planar stack configuration should be modeled in Phase I?
Rectangular or circularCo-, counter-, or cross-flow
What design details (e.g., seals, manifolds, insulation) should be included in the Phase I modeling effort?What size (kW) stack should we consider?
17
Manufacturing Model Performance/Structural Module
What choices affecting both cost and performance should we analyze?
For example, we could consider the impact of layer thickness on system power and thermal stresses.
LayerThickness
ProcessingSteps
OhmicLosses
StackSize
ThermalStresses
CrackingFailures
PartYield
MaterialCost
ProcessCost
Area-SpecificResistivity
PowerDensity
InputsInputs Performance/Structural ModulePerformance/Structural Module Cost ModelCost Model
$ kW
18
Manufacturing Model Use of Performance/Structural Module
In addition, the Performance/Structural Module could be used for stand-alone simulations to evaluate the sensitivity of particular material or process parameters.
ContactResistance
ContactArea
PowerDensity
SurfaceFinish
Performance/Structural ModulePerformance/Structural Module
ContactResistance
PressureVariations
PowerDensity
ThicknessVariations
MechanicalStresses
MEAWarping
CrackingFailures
Oven TemperatureVariations
What design parameters, material properties, or manufacturing conditions are of interest for analysis, either in Phase I or II?
19
Manufacturing Model Parameters Baseline Case Design
As a basis for Phase I, we will use an anode supported design.
Anode/Electrolyte/CathodeAnode/Electrolyte/CathodeAnode/Electrolyte/Cathode One-Half Interconnect LayerOneOne--Half Interconnect LayerHalf Interconnect Layer
Ni Cermet Anode700 µm
8YSZ & LSMCathode 50 µm
Y-stabilized ZrO2Electrolyte
10 µm
Ferritic Stainless Steel 4 mm
2 mm
We will only assess the stack costs in this phase. We also considered inclusion of reforming layers or materials in the stack, but have insufficient design information in this Phase of work.
20
Manufacturing Model Parameters Performance Parameters
We propose using the following set of operating parameters for the stack.
ParameterParameter Value/Range
• Cell voltage
• Power Density• Composition of the reactant
streams• Gas inlet temperatures• Fuel utilization • Cathode stoichiometry
• Cell voltage
• Power Density• Composition of the reactant
streams• Gas inlet temperatures• Fuel utilization • Cathode stoichiometry
• 0.7 V• 500 mW/cm2 (not reactant limited)• Anode: reformate; Cathode: air
• 650°C at the Anode and Cathode• ~ 50 %• ~ 5, adjusted to effect an exit temperature of
800°C.
• 0.7 V• 500 mW/cm2 (not reactant limited)• Anode: reformate; Cathode: air
• 650°C at the Anode and Cathode• ~ 50 %• ~ 5, adjusted to effect an exit temperature of
800°C.
Value/Range
The performance model will calculate the current distribution over the electrode, the average power density, and the actual fuel utilization.
21
Manufacturing Model Parameters Impact of Layer Thickness
We will look at the trade-offs between layer thickness and their impact on performance and cost. The latter impacted by material quantities and yield.
Layer NominalThickness (µm) Remark
Anode
Material
700
• Minimize thickness to reduce material weight and resistance
• Impact of thickness on strength and MEA stress
Electrolyte
Ni-YSZ
10• Barrier properties vs thickness critical• Impact of coating technology and thickness
on defects
Cathode
YSZ
50 • Coating technologies
Interconnect
YSZ- LSM
• Roll form technique used in baseline study Metal 4300
As part of this effort we will look at the impact of the attributes of various process technologies on each layer, types of defects, and number of defects. It will be critical to find relationships between defects, materials, and processes.
22
Manufacturing Model Parameters Economies of Scale
We will consider how production volume impacts cost ($/kW).
Assumptions5 kW unit sizeunit operations are automated to achieve uniformity and maximizeyieldincreasing volume can change equipment scale, speed, material logistics in the process, and automation of assembly
ParametersDays per weekShifts per day
Commercialization (Volume) StepsProduction PrototypesMarket EntryMarket Penetration
Our 1999 study was made assuming 250 MW, however, we have not fixed the volume steps at this time for this project.
23
Manufacturing Model Parameters Process Flow Options
We will look at a multi-fired process flow option in Phase I.
Co-fired Process FlowCoCo--fired Process Flow Multi-fired Process FlowMultiMulti--ffired Process Flow ired Process Flowfired Process Flow
Tape Cast
AnodePowder Prep
VacuumPlasmaSpray
ElectrolyteSmall Powder
Prep
ScreenPrint
CathodeSmall Powder
Prep
Sinter in Air1400C Sinter in Air
ProgressiveRolling of
Interconnect
ShearInterconnect
VacuumPlasmaSpray
SlurrySpray
ScreenPrint
Slurry Spray
Slip Cast
Multi-FiredProcess
Finish Edges
BrazePaint Braze
ontoInterconnect
Blanking /Slicing
QC LeakCheck
Interconnect
Fabrication
Electrolyte CathodeAnode
Stack Assembly
Co-Fired Process Flow
AnodePowder Prep
CathodeSmall Powder
Prep
ElectrolyteSmall
Powder Prep
Tape Cast
Tape Cast
Tape Cast
Blanking /Slicing
StackCalendar
Dual AtmSinter
Diamond GrindEdges
Slip Cast
Slip Cast
Slip Cast
Roll Calendar
ShearInterconnnect
ProgressiveRolling of
Interconnect
Note: Alternative production processes appear in gray to thebottom of actual production processes assumed
Blanking /Slicing
Paint Brazeonto
InterconnectBraze
QC LeakCheck
Interconnect
Fabrication
Electrolyte
Cathode
AnodeStack Assembly
• Individually tape-cast layers
• Laminated together
• Co-fired in one step
• Individually tape-cast layers
• Laminated together
• Co-fired in one step
• Tape-cast anode layer
• Electrolyte and cathode layers applied by coatings
• Sequential firing steps
• Tape-cast anode layer
• Electrolyte and cathode layers applied by coatings
• Sequential firing steps
24
Manufacturing Model Discussion Non-Technical Issues
In this phase, we will use generic data, however, for Phase 2 we will have to develop a protocol(s) for protection of proprietary information with participating teams.
Issues:
Protection of individual SECA team proprietary informationSecurity of User InterfaceAccess to modelAccess to process and equipment data and specifications
Protection of individual SECA team proprietary informationSecurity of User InterfaceAccess to modelAccess to process and equipment data and specifications
Java based UserInterfaceUserUser
InterfaceInterface
Activity-Based CostModel
Activity-Based CostActivity-Based CostModelModel
ManufacturingProcess DatabaseManufacturingManufacturing
Process DatabaseProcess Database
ManufacturingProcess Flow
ManufacturingProcess Flow
Thermal SprayThermal Spray
Tape CastingTape Casting
SinteringSintering
Materials DatabaseMaterials DatabaseMaterials Database
LSMLSM
YSZYSZ
316 Stainless Steel316 Stainless Steel
SOFC ScenarioCompiler Module
SOFC ScenarioSOFC ScenarioCompiler ModuleCompiler Module
Issues:Issues:
25
Next Steps Phase I
We expect Phase I to be completed in approximately 3-6 months.
Modify Model and Analyze Selected Scenarios and IssuesLayer thickness and processesEconomies of scale
Discuss results with SECA teamsDevelop plans for Phase 2
Phase I Final Report
26
Activities for Phase I TIAX Team Members
The TIAX core team consists of five members whose backgrounds are particularly appropriate to this project.
Staff Email Telephone
Yong Yang [email protected] 617-498-6282
Eric Carlson
Project Input
[email protected] 617-498-5903
Chandler Fulton
Principal Investigator
[email protected] 617-498-5926
Suresh Sriramulu
System modeling
[email protected] 617-498-6242Fuel cell technology
Manufacturing model
