Skip to main content

Automated mass-manufacturing and quality assurance of Solid Oxide Fuel Cell stacks

Periodic Reporting for period 1 - qSOFC (Automated mass-manufacturing and quality assurance of Solid Oxide Fuel Cell stacks)

Reporting period: 2017-02-01 to 2018-07-31

qSOFC project combines leading European companies and research centres in stack manufacturing value-chain with two companies specialized in production automation and quality assurance to optimize the current stack manufacturing processes for mass production. Currently the state-of-the-art SOFC system capital expenditure (capex) is 7000…8000 €/kW of which stack is the single most expensive component. The qSOFC project focuses on SOFC stack cost reduction and quality improvement by replacing manual labour in all key parts of the stack manufacturing process with automated manufacturing and quality control. This will lead to stack cost of 1000 €/kW and create a further cost reduction potential down to 500 €/kW at mass production (2000 MW/year). During the qSOFC project, key steps in cell and interconnect manufacturing and quality assurance will be optimized to enable mass-manufacturing. This will include development and validation of high-speed cell-manufacturing process, automated 3D machine vision inspection method to detect defects in cell manufacturing and automated leak-tightness detection of laser-welded/brazed interconnect-assemblies.

The project is based on the products of its' industrial partners in stack-manufacturing value-chain (ElringKlinger, Elcogen AS, Elcogen Oy, Sandvik) and motivated by their interest to further ready their products into mass-manufacturing market. Two companies specialized in production automation and quality control (Müko, HaikuTech) provide their expertise to the project. The two research centres (VTT, ENEA) support these companies with their scientific background and validate the produced cells, interconnects and stacks. Effective exploitation and dissemination of resulting improved products, services, and know-how is a natural purpose of each partner and these actions are boosted by this project. This makes project results available also for other parties and increases competitiveness of the European fuel cell industry.
During the first reporting period, the first prototype of the automated machine vision inspection system to be used in cell quality assurance has been developed. A functional prototype has been designed to detect five different categories of defects performing 100% inspection of the green and fired cells in 10 s per cell. The prototype was first trained with faulty cells in the cell manufacturing process. The prototype showed real-time detection accuracy of >90% for fibers and holes and 65% for dark spots and cracks in cell layers, which is a very promising first result. The prototype will be further developed during the project and the improved version will be validated again at Elcogen AS production line.

Scaling up cell manufacturing while lowering cost-level requires higher manufacturing speeds. HaikuTech and Elcogen AS have carried out development of tape-casting and screen-printing processes with a goal of reaching at least 1 m/min tape casting speed and <10 s/cell screen printing speed. Both development processes have required more efforts than originally thought, but the work is progressing well and the latest experiments show that cells through the “fast” route show close to similar electrochemical performance compared to the conventional process. That said, still more work is needed to optimize the pastes and slurries for high-speed manufacturing. During this development process, all of the used materials are also analysed for compliance with REACH as well as other relevant regulations. Some of the regulated materials have already been substituted with non-regulated materials and this work will continue.

For quality assurance of complete interconnect-assemblies, ElringKlinger and MüKo have been developing several approaches, including leakage testing, alignment and traceability-solutions. By carrying out leak-testing of complete interconnects and using data matrix code marking on the interconnects, valuable data of alignment and leak tightness, amongst other process conditions, can be obtained and correlated to gain better understanding of the manufacturing tolerance requirements for interconnect-assemblies. Optimization of interconnect manufacturing progress was started by carrying out a failure mode and effect analysis for the whole manufacturing process. This allowed to locate critical process steps and subsequently to focus quality assurance efforts in improving these. Forming of pre-coated steel substrate material has been investigated and a number of different processing options have been studied and compared against specification derived from Elcogen Oy stack assembly. Furthermore, it was decided necessary to carry out oxidation and area-specific-resistivity (ASR) tests for samples manufactured through different process routes in order to assure that the best route from the manufacturing point-of-view will still yield good electrochemical performance in stack.

For stack-level manufacturing and quality assurance, a new production quality control tool for stack pre-compression has been designed and made operational at Elcogen Oy. The tool has been shown to yield more homogenious stacks and therefore better quality. Testing different stack conditioning procedures aims at reducing the CAPEX and OPEX in stack manufacturing by simplifying the conditioning process which is carried out to manufactured stacks at factory before shipping to customers. Currently ten stacks out of 30 have been tested at VTT and the tests are continuing until the very end of the project. At the moment, the tested number of stacks is still in too low for a clear indication if there is a statistical correlation between any conditioning process and stack performance, but it is clear that all of the conditioning processes result in usable stacks with fairly small differences. As the number of tested stacks increase, possible more subtle differences in performance will be easier to note by statistical means and this work will continue u
Cell quality assurance work will continue by modifications to the automated machine vision inspection system and fine-tuning and training of the defect-detection algorithms. The evolved prototype will be again tested first at HaikuTech and then at Elcogen AS cell manufacturing line. It is expected that the detection accuracy for various defect-categories will be even higher than with the first generation prototype. Development of slurry and paste compositions for higher manufacturing speeds will continue to reach the target speeds for mass manufacturing.

The work on interconnect quality assurance will continue by area specific resistance (ASR) testing of coated steel substrates manufactured through different processing routes. This work will be continued by development of automated leak detection protocols for complete interconnect assemblies. Finally, interconnect will be manufactured through the optimized processing routes for the final validation stack test.

On stack-level, the evaluation of conditioning procedures will continue and as more data is gathered, the three different conditioning options can be evaluated in more detail. Towards the end of the project, possible performance differences in stacks manufactured through each route will be weighed against savings in CAPEX and OPEX made possible by the streamlined conditioning processes.