Spectrum Instrumentation has reported that ZBT has an ongoing research project to improve the design of hydrogen fuel cells, using 8-channel digitizers M2i.4652 from Spectrum to analyze the performance of different fuel cells on a test bench, providing real-world data to improve the virtual cell model.
Ing. Sönke Gössling, Group Leader for Fuel Cell Systems Simulations and Controls, explained: “The computer model of the fuel cell is very sophisticated such that we can adjust the many variables that affect its performance to see what changes provide performance enhancements. However, these are just theoretical, so the test bench enables us to see how changes in real-world parameters affect performance. We quickly realized that the data capture of the rate of every second was not providing us with the level of fine detail that we needed
“We now use three Spectrum digitizers that dramatically improved our data rate capture to three mega-samples per second as well as having twenty simultaneous channels of data. This enables us to analyze the dynamic step changes as well as analyze superimposed high frequencies at an incredible level of detail. The synchronization of the cards with each other and the connection to the test bench environment were intuitive and free of complications. The performance and quality of the cards is first class and they have worked perfectly from day one.”
Ing. Gössling said: “Validating predictions with real-world results is a vital part of the scientific method and will help us improving fuel cell design to meet our target of significantly reducing the costs of fuel cells, which will open up adoption by becoming more economical and competitive.”
The measurements make it possible to get an insight into the processes inside the fuel cell. They answer the question of how the processes are distributed with which dynamics within the fuel cell. This is of crucial importance, for example, to avoid local undersupply in dynamic operation or to optimize the operating conditions in a focused manner.
If the computer model can be validated with the data, the reliability of the predictions of the model increases in general. As a result, development and optimization processes can increasingly be carried out virtually, which promises a major cost and time advantage.