Proton Exchange Membrane (PEM) fuel cell technology, a low-emissions and low temperature power source receiving much attention for its efficiency, will need to progress from low-volume production to high-volume within the course of the next decade. To successfully achieve this transition, development of a fully-functional fuel cell automatic stack assembly robotic station must be completed, and the fuel cell stack must be redesigned to become more readily automated. This document outlines the progression of research completed in the development of this automatic fuel cell stack assembly station and the subsequent stack redesign.

Numerous patents exist pertaining to equipment and processes for fuel cell assembly as well as a stack features to aid in manufacture or assembly. However, most of this is focused upon proper compression of the membrane material, with little thought given to overall assembly and throughput. Therefore, there is a need for more contributions to stack manufacture and assembly.

With the intention of developing the aforementioned manufacturing workcell, three individual robotic fuel cell assembly workstations have been designed and constructed. The first iteration successfully built a five-cell stack. The second iteration incorporated numerous improvements, including overlapping work envelopes, elimination of a shuttle cart, software synchronization, fewer axes, and a better end effector. Consequentially, the second workcell achieved a four-fold improvement in cycle time over the previous iteration. The third iteration improved reliability of the overall system via humidity control, added endplates via an endplate feeder and gripper, and completed stack fastening via a bolt feeder and torque-control end effector.

During the course of working with the third robotic workcell, it was recognized that current stacks contain many features that impair automation efforts. Multiple changes were made to several of the stacks to improve the ability to automate the assembly of each stack, and the results of these modifications were tested in the third workcell. These changes had a dramatic impact on the reliability of the automation workcell.

Meanwhile, the nature and limitations of two purchased types of fuel cell stacks were examined. The first of these stacks, TekStak 5-cell educational fuel cell stacks, were tested for component tolerances, leak testing, and performance testing. These tests revealed a number of potential avenues for stack improvement. Afterwards, a Plug Power® GenCore® fuel cell stack was tested for overall dimensional coordinates, leak testing, and component tolerances. These tests revealed additional insights regarding the manufacture and assembly of industrially-relevant stacks.

Finally, using the design-for-manufacture/assembly (DFMA) guidelines set forth in Product Design for Manufacture and Assembly by Geoffrey Boothroyd, et al, a DMFA analysis was conducted to determine additional potential areas for stack improvement in order to reduce assembly time and overall cost. This allowed for a theoretical stack redesign, in order to bring the theory and practice of fuel cell manufacture / assembly more in line with one another.

Based upon the knowledge gained, a set of guidelines for stack DFMA can be recommended. Future work includes characterization of stack failure modes, the tracing of these failure modes to assignable causes, the implementation of quality control procedures to avoid stack failures to reduce or eliminate the need for stack testing.