Cost Modeling of SOFC Technology

Transcription

1 Cost Modeling of SOFC Technology Eric J Carlson Suresh Sriramulu Peter Teagan Yong Yang First International Conference on Fuel Cell Development and Deployment University of Connecticut Storrs, Connecticut March 10, Acorn Park Cambridge, Massachusetts carlsone@tiaxllccom

2 Summary Updates to a 1999 SOFC cost model resulted in less than a ten percent increase in the cost of the stack on a kilowatt basis Model Total Cost ($/kw) Co-Fired Multi-Fired Drivers for Lower Cost Lower YSZ material cost Drivers for Higher Cost More interconnect material Additional QC steps and equipment Overall process yield assumptions Slightly lower power density for the baseline case SOFC Cost Modeling Presentation_UConn Conference_Mar04ppt 1

3 Project Overview For the SECA Core Technology Program (CTP), we updated a 1999 SOFC cost projection Cost and performance/mechanical models linked to capture the influence of design, performance, and mechanical limitations on cost Assessed the impact of manufacturing issues (eg, tolerances and quality control) on cost Model used to assess the impact of manufacturing volumes on cost We solicited inputs from the SECA industrial teams and the CTP participants SOFC Cost Modeling Presentation_UConn Conference_Mar04ppt 2

4 Modeling Approach Manufacturing Cost The model uses a set of databases to calculate cost for defined production/process flow scenarios Inputs Design Performance Parameters Processes and Process Flow Production Scenarios Calculation Engine (Activity-Based) Process flow Equipment options Outputs (Results) Tables Graphs Crystal Ball Analyses Manufacturing Costs Material Cost Database Material Properties Database Purchased Components Formulation Database Process Database Capital Equipment Database Labor Real estate Overhead Vendors Cost of material Density Particle size distribution Surface area Stress vs probability of failure Cost Specifications Anode Cathode Electrolyte Interconnects Equipment process data Throughput Size limit Automation Scrap Yield Cost vs product volume SOFC Cost Modeling Presentation_UConn Conference_Mar04ppt 3

5 Modeling Approach Performance Thermal and Stress A performance-thermal-mechanical model developed for NETL was used to estimate power density and stress as a function of layer thickness Performance Model Material Failure Data + Stress Probability of Failure (%) Electrodes Electrolyte Stress (MPa) Parameters from ORNL, Edgar Lara-Curzio Power Density Stress Related Yield Losses SOFC Cost Modeling Presentation_UConn Conference_Mar04ppt 4

6 Model Assumptions The 2010 SECA goals target a system manufacturing cost of $400/kW This project focused on the stack materials only Only the electrochemical (anode, cathode, and electrolyte) and interconnect materials are considered in this model The interconnect cost does not include a coating Factory costs were estimated Corporate overhead, profit, and installation costs were not included High volume production was assumed for the baseline cost estimate (total of 250 MW with 5 kw stack as basic unit) SOFC Cost Modeling Presentation_UConn Conference_Mar04ppt 5

7 Results 2003 Baseline Material Costs (Area Basis) On an area basis, the 2003 model material cost decreased, largely driven by the reduced electrolyte (YSZ) cost $250 $200 Co-Fired 1999 Multi-Fired 2003 Multi-Fired Material Costs Costs $250 $200 Multi-Fired $/m2 $150 $100 $/m 2 $150 $100 $50 $50 $- Anode Cathode Electrolyte Interconnect $- Anode Cathode Electrolyte Interconnect Material Cost ($/m 2 ) Co-Fired Multi-Fired 1999 Model Model SOFC Cost Modeling Presentation_UConn Conference_Mar04ppt 6

8 Results 2003 Baseline Process Costs by Layer (Area Basis) Process costs increased by 60-75% because of added QC steps, the final assembly step, and reduced yields Process Costs Costs $120 Co-Fired $140 Multi-Fired $100 $120 $80 $100 $/m2 $60 $40 $/m2 $80 $60 $40 $20 $20 $- Anode Cathode Electrolyte Interconnect Fabrication $- Anode Cathode Electrolyte Interconnect Fabrication 1999 Multi-Fired 2003 Multi-Fired Processes Cost ($/m 2 ) Co-Fired Multi-Fired 1999 Model $82 $ Model $143 $169 SOFC Cost Modeling Presentation_UConn Conference_Mar04ppt 7

9 Results Anode cost is large because of the materials costs, while the interconnects are massive Material cost ($/m 2 ) Cost ($/m 2 ) Materials Total Interconnect 45% Anode 48% Anode $126 Cathode $15 Electrolyte $6 $136 $22 $12 Interconnect $119 $138 Electrolyte 2% Cathode 5% Fabrication $126 Total $266 $434 Materials represent approximately 60% of the stack cost SOFC Cost Modeling Presentation_UConn Conference_Mar04ppt 8

10 Results 2003 Cost Vs 1999 Cost In 2003 lower material costs partially offset the increases in process cost resulting in similar $/kw cost and $/m 2 costs with the previous study $ / kw $100 $90 $80 $70 $60 $50 $40 $30 $20 $10 $0 Total Cost ($/kw) Total Cost ($/m 2 ) $ / m2 $450 $400 $350 $300 $250 $200 $150 $100 $50 $0 Co-Fired Multi-Fired Co-Fired Multi-Fired Co-Fired Multi-Fired Co-Fired Multi-Fired Material Process SOFC Cost Modeling Presentation_UConn Conference_Mar04ppt 9

11 Results Sensitivity Analysis Base Case Unit cell cost per kilowatt is most sensitive to the thickness of each EEA layer and YSZ price Sensitivity Chart Target Forecast: Multi-Fired Metal Planar Cost ($/kw) Anode Layer Thickness (um) 59 Electrolyte Layer Thickness (um) 58 Cathode Layer Thickness (um) 30 8 mol% YSZ 24 SS430 sheet 21 Nickel Sintering Furnace 09 Sintering Cycle Time (min) 07 Auxiliary Equipment Cost 06 Equipment Installation Cost Measured by Rank Correlation The electrolyte cost is small, but its thickness has a large impact on power density SOFC Cost Modeling Presentation_UConn Conference_Mar04ppt 10

12 Results Stack Model Achieving high power densities is critical to lower stack costs 120 Multi-Fired Cost vs Power Density Cost ($/kw) Anode Cathode Layer Thickness (µm) Electrolyte 5~20 Status Fixed Fixed Vary Power Density (mw/cm 2 ) SOFC Cost Modeling Presentation_UConn Conference_Mar04ppt 11

13 Results Quality Control Assembled stack cost will be highly sensitive to the percentage of defective EEAs Stack Cost ($/kw) $700 $600 $500 $400 $300 $200 $100 $0 00% 05% 10% 15% 20% Percentage of Defective EEAs For example, for the MF process, a 1% defect level could increase the stack cost from $92/kW to $278/kW SOFC Cost Modeling Presentation_UConn Conference_Mar04ppt 12

14 Results Economies of Scale The stack cost decreases by 80%, driven by more efficient processing, as production volume increases Economies of Scale (Multi-Fired Process) $ $ Stack Cost ($kw) $300 $200 MW Reduction Material $73 $56 13x Process $385 $6 107x Stack Cost % of Reduction $ $ Annual Production Volume (MW) SOFC Cost Modeling Presentation_UConn Conference_Mar04ppt 13

15 Conclusions Increasing power density will be critical to achievement of low stack cost since materials represent approximately 60% of the cost Quality control of the repeat units (electrode electrolyte assemblies) will be critical to stack yield and cost Increasing production volume 50-fold from 5MW to 250MW decreased process costs resulting in an 80% reduction in stack cost SOFC Cost Modeling Presentation_UConn Conference_Mar04ppt 14

16 Acknowledgements Thank You! This project was funded under DOE contract: DE-FC26-02NT41568 Final Report available in April We would like to thank: Shawna Toth, Donald Collins, and Wayne Surdoval of the National Energy Technology Laboratory and the SECA Core Technology and Industry Teams SOFC Cost Modeling Presentation_UConn Conference_Mar04ppt 15