Zero Emission Fuels (ZEF B.V.) is a start-up working to build a sustainable methanol micro-plant. Carbon dioxide and water are obtained from atmosphere. Hydrogen is obtained from captured water through alkaline electrolysis and the hydrogen is used for methanol production. The alkaline electrolyser should run at 50 bar and 90°C with 30% KOH as the electrolyte. It is a multi-cell design and the electricity supplied to the electrolyser leaks within the system through the electrolyte. This is the energy supplied to the system not being used for splitting water into hydrogen and oxygen. Leaking current and flow scale with the length and the radius of any channel in the system and contradict each other in terms of requirements.
This study focuses on characterising the electrolyser to validate a modified version of the existing Matlab model at a channel in the system for two different dimensions of the channel. The Matlab model predicts
the flow and the leaking currents in the system. Estimating the flow in the system is essential to validate the model and using the model, the leaking currents are estimated. Experiments were performed on a single cell version of the electrolyser at atmospheric pressure to obtain the flow and thermal characteristics. Comsol simulations were ran on the single cell system to support the Matlab results.
The heating curve of the system was obtained at four different points in the system to check for the necessity of insulation. The system went up to 56°C at the hottest point in the system after 2 hours. This confirmed for the necessity of insulation. The diameter of the bubbles at the electrode were measured and compared to the estimated diameter in Matlab. They were 34% off. The flow part of the experiments were done at a lower current density to simulate 50 bar flow at 1 bar. The flow rate of the electrolyte was estimated using high speed cameras and tracking fluorescent particles at one of the channels in the system. The values of the flow rate from Comsol, Matlab and Comsol were 25% away from each other whereas Matlab and experiments were 65% away from each other. The mass flow rate predicted by Comsol was in between Matlab and experiments. The absence of a temperature network to estimate temperature at every part of the system, the approximate geometry and ignoring smaller resistances to flow in the Matlab model are responsible for the offset.
Leaking currents were estimated in the single and multi-cell system using Matlab. The power lost due to leaking currents were higher in the multi-cell system as compared to the single cell system. The reduction in leaking currents by using gas slugs was estimated and the current design achieves the electrical efficiency of 99%. Recommendations were to make the channel diameter larger and remove the part of the channel that is in the cell. Further modifications to the Matlab and Comsol model were suggested to improve the prediction. Experiments related to multi-cell setup have to be done to estimate the actual electrical efficiency of the system and validate the Matlab model.