Membrane Humidifier Research Articles

CFD simulation of the interaction between water droplets on the GDL of a low temperature PEM FC in conditions of air cross-flow

Abstract Droplet dynamics in low temperature proton exchange membrane fuel cells (LT-PEMFCs) are significantly important with respect to water management and removal. Three main phenomena are linked to water generation/transport in a LT-PEMFC: H2O production on the cathode side, electro-osmotic drag (anode to cathode) and back diffusion (cathode to anode). The excess/lack of water, depending on membrane humidity, strongly affects LT-PEMFC performance, involving its instability. In this work CFD simulation of droplets was performed when they were deposited in tandem on the gas diffusion layer (GDL) and subjected to 15 m/s parallel air flow. The volume of fluid (VOF) method was adopted and droplets with diameters that range from 0.3 mm to 0.8 mm were considered. Simulations demonstrate that deformations, oscillations and sliding of droplets were affected by their relative size and position. Results can contribute to define predictive models of water droplet dynamics on the GDL surface and their detachment towards the FC gas channel.

Comprehensive analysis of polysulfone membrane humidifier in hydrogen fuel cell vehicles: Experimental and theoretical approaches

This study performed a comprehensive experimental and theoretical analysis of a shell-and-tube type gas-to-gas humidifier module in a 95 kW class hydrogen fuel cell vehicle. The humidifier module features a polysulfone hollow fiber membrane characterized by its hydrophilic and non-fluorinated properties. Experimental investigations were conducted to understand the impacts of the inlet temperature, relative humidity, and temperature differences on heat and mass transfer rates, and the approach dew point temperature (ADT). The results showed that the ADT increased from 12.42 to 15.89 °C with a temperature difference of 15 °C between the shell and tube inlet sides, as the flow rate increased from 1000 to 3400 L/min, with the highest performance observed at lower flow rates. Simultaneously, a detailed theoretical model was developed, incorporating a new dual-mode sorption approach that accounts for temperature dependency within the solution-diffusion mechanism across the membrane. The developed model was in good agreement with the experimental results, with a relative error of less than 5 %. Furthermore, the dynamic model accurately predicts the response of the humidifier module to step changes in feed flow rates, simulating load variations in the fuel cell vehicle. This provides valuable insights into the humidifier's adaptability and responsiveness, which are crucial for optimizing its performance under various operating conditions.

The Effect of Cation Contamination on the Water Transport for PEM Water Electrolysis

Large-scale hydrogen production by PEM water electrolysis (PEMWE) in the energy sector requires a deep understanding of the system behaviour in failure cases and possible recovery procedures. This includes the purification of the anolyte feed water because the contamination with ions can cause the PEMWE stack to breakdown.[1] For instance, the presence of iron trace cations or other metal cations in the feed water leads to a performance loss due to higher ohmic resistance of the cell and higher iR-free voltages.[2] [3] Moreover, cation contamination might influence the water transport in a PEMWE cell. In principle, the transfer of the water molecules to the cathode is proportional to the proton flux from the cathode to the anode. The proportionality factor is defined as water transport coefficient and is also temperature dependent.[4] In case of the mono- and multivalent cations as trace elements in the feed water, the water loading of the membrane, which is expressed by water molecules per sulfonic acid group of the membrane, λ, decreases over time and leads to successive increase in the ohmic resistance. [5] However, the influence of cation contamination on the water transport in PEMWE is poorly understood to date.In this work, we investigated the effect of cation contamination in the anode feed water on the water crossover for PEMWE. Cation contamination experiments were carried out at a constant current density by adding different concentrations such as K2SO4, Na2SO4, etc. to the anode feed water. All electrochemical measurements were carried out in an in-house test bench equipped with a single cell of 5 cm2 geometric electrode area and potentiostat with booster. A commercially available catalyst coated perfluorosulfonic acid membranes (loading of 0.3 mg cm-2 geo Pt/C and 1 mgcm-2 geo IrOx) was used. The anode and cathode compartments were kept at atmospheric pressure and cell temperature of 80°C. To determine the water crossover from anode to cathode at different current densities, the cathode exhaust was cooled in a heat exchanger and measured by a gravimetric method. [4] Firstly, a conditioning procedure was performed with purified feed water until steady cell performance was achieved. The cell performance was evaluated using polarization curves, electrochemical impedance spectroscopy (EIS) and water crossover measurements. The last two are of particular interest, because they provide information about membrane humidification. The Tafel slope and mass transport overvoltage were established from the polarization curves and high frequency resistance (HFR). We observed a significant rise of the cell voltage by adding cations to the anode water feed within few hours. The analysis of Tafel slope and EIS data also reveal an increase of ohmic and mass transport resistances. In these experiments, we were able to correlate the HFR results with the water crossover as a function of the nature of the cation and its concentrations. Our data shows that the specific water crossover per current density decreases as a consequence of cation contamination. The switch back to purified feed water allows us to monitor the dynamics of the performance recovery process obtained from EIS and water crossover measurements. Additional measurements of the polarization curves and HFR were used to distinguish between reversible and irreversible degradation processes due to the cation contamination.Altogether, we present the effect of cation contamination on the water transport processes and provide deeper insights into the mass transport and performance losses for PEMWE.

Technical evaluation of electrochemical separation of hydrogen from a natural gas/hydrogen mixture

In this work, we investigate the electrochemical separation of hydrogen from a 10% H2/methane natural gas blend using a single cell experimental setup for electrochemical hydrogen pumping (EHP). Results show that the overall efficiency for hydrogen separation is high, ranging from about 87% at 0.2 A/cm2 to 75% at 0.6 A/cm2. In addition, we develop a zero-dimensional model for the hydrogen separation process including the membrane permeation of H2 as well as CH4, N2, and CO2 as the main components in natural gas. We derive analytical expressions for the permeability of these gases as a function of temperature and membrane humidity using available permeability data. Although the presence of dilutive inert gases poses a challenge, the modelled data indicate that it should be possible to produce fuel cell quality hydrogen using a single hydrogen separation stage. Importantly, this indicates technological feasibility of distributing hydrogen for use in fuel cells through existing natural gas infrastructure. Extension of this research to combined hydrogen separation and compression within the same EHP unit is recommended as future work.

Experimental study on the mass transfer permeability of hollow fiber membranes for a humidifier in a proton exchange membrane fuel cell

A membrane humidifier is a device used to increase proton conductivity in a fuel cell system, enhancing the performance and longevity of the fuel cell. In this study, a hollow fiber membrane is tested under various operating conditions to determine the membrane humidifier performance by estimating the amount of water vapor through the membrane. The results show that the mass transfer rate increases with increasing temperature, flow rate, and relative humidity while it decreases with pressure and membrane thickness. Particularly, temperature affects the mass transfer rate the most. When the temperature is increased from 50°C to 80°C, the performance improves by approximately 4.5 times. In contrast, the flow rate affects it the least; an approximately 2.5 times increase in the flow rate results in a 16% improvement in performance. Furthermore, the counterflow consistently exhibits greater efficiency than the co-flow. Moreover, the water vapor permeability through the membrane is assessed. The membrane permeability increases with decreasing membrane thickness and pressure and increasing temperature and inlet relative humidity. The findings indicate that 1°C rise in temperature leads to an increase of approximately 5% in permeability and approximately 19% increase in pressure, resulting in an approximately 15% decrease in permeability.

Adaptive extended kalman filter for PEMFC membrane water content estimation

Proton Exchange Membrane Fuel Cells are a favorite technology for decarbonizing the transportation sector. However, their large-scale democratization is hampered by their high cost compounded by their unsatisfactory lifespan. To anticipate potential degradation while keeping improving performance, it is essential to maintain an acceptable humidity range inside the cells, especially at the membrane level. However, membrane humidity level is not directly measurable, alternative techniques must be considered to recover this key variable. Here, we develop a real-time software sensor of the membrane water content at the fuel cell’s heart. We build a model describing the membrane water balance, electrochemical behavior, and species mass balance. We then reduce the model and perform an Adaptive Extended Kalman Filter. We perform sensitivity analyses in both steady-state and transient conditions. We validate the filter on a “Worldwide Harmonized Light Vehicles Test Cycles” test procedure. Finally, we obtain a fast and accurate model-based software sensor.

How ozone traces degrade polyimide humidifier membranes for fuel cells

Humidification of polymer electrolyte membranes in fuel cells is essential for high proton conductivity and lifetime, therefore often membrane humidifier modules are used. We report about the degradation of polyimide humidifier membranes under the influence of airborne ozone traces: during operation we tracked the membranes humidifier performance in 5 modules for up to 1000 h with trace levels of ozone (100 ppb) and conducted characterization tests at 200 h intervals. Operating the humidifier with ozone resulted in a linear decrease in the membrane's ability to transfer moisture over time. Moreover, the glass transition temperature of the membrane material decreases linearly with longer exposure to ozone, while the mechanical strength in terms of breaking force and elongation at break decreases too. Infrared spectra of the tested fibers showed no changes. The reduced water vapor flux would limit fuel cell performance, while the reduced mechanical properties of the membrane can lead to rupture.

Comparison analysis of entropy and entransy under various conditions of a hollow fiber membrane humidifier (HFMH)

The hollow fiber membrane heat and moisture transfer model is converted equivalently into a series of parallel-plate membranes stacked along the z-axis to simplify the model. A detailed modeling process is described based on the concept of finite difference method. The comparison of experimental values with calculated values ensures the validity of the model. Subsequently, an entransy dissipation model for the heat and moisture transfer process in a hollow fiber membrane humidifier (HFMH) is developed by analogizing heat and moisture conduction to electrical conduction. The variation of heat/moisture transfer capability within the HFMH is investigated from the perspective of “heat/moisture potential energy”. In addition, an entropy generation model for the HFMH is established to study the irreversible losses in the heat and moisture transfer process caused by irreversibility. The performance, entropy generation rate, and entransy dissipation rate of the HFMH under different conditions are comparatively analyzed. It can be found that the heat/moisture transfer performance of the humidifier can be improved by adjusting the operating conditions to increase the heat/moisture transfer driving force. However, the irreversible losses increase due to a higher heat/moisture transfer energy grade. The sensible and latent effectiveness can be enhanced by 20.9%–85.2 %, while the equivalent resistance can be reduced by 16.2%–42.0 % via optimizing the geometry structures and membrane properties. In the hot-dry region, the heat and moisture transfer increase by 25.9%–89.8 %, while the equivalent resistances decrease by 20.8%–66.0 %. In addition, the minimum equivalent resistances correspond to the optimal HFMH performance. The entransy analysis avoids the “paradox” of entropy analysis.

Flexible cellulose nanofibers/MXene bilayer membrane humidity sensor with a synergistic effect of force and hygroscopic expansion

A double layer membrane humidity sensor was designed with a synergistic effect of force and hygroscopic expansion, capitalizing on the high sensitivity demonstrated by MXene material. This sensitivity was achieved through the modulation of resistance change associated with layer spacing of MXene. The double layer structure of sensor was made by vacuum filtration, with an inner layer of MXene for conductivity and an outer layer of cellulose nanofibers (CNF) for moisture absorption, which is rooted in the MXene through the CNFs, similar to the biomimetic structure of hairs rooted in the skin. The synergistic effect of humidity sensitivity is achieved by trapping and dissociating water molecules through the CNF layer generating a change in MXene stress and disrupting the conductive pathway of MXene upon swelling. The CNF/MXene humidity sensor showed a resistance change rate of 117.93 % from 11 % RH to 98 % RH, accompanied by a response time of 319s and a recovery time of 132s. Stable sensing performance is maintained even after 5 times adsorption and desorption cycling, prolonged exposure (7 days) or bending and folding. CNF/MXene bilayer membrane also has good electrical heating performance, reaching about 80 °C at 7V and maintaining stable performance under 5 cycles of heating and 10 min of long heating, respectively. The CNF/MXene membrane humidity sensor can monitor human respiration monitoring, fingertip humidity, pipeline leakage and artifact dehumidification, and combined with the function of electric heating can realize the cycle of dehumidification and humidity monitoring, which has a wide range of application prospects in flexible equipment.

Novel electrochemical and thermodynamic conditioning approaches and their evaluation for open cathode PEM-FC stacks

The Air-cooled and open cathode proton exchange fuel cell (PEMFC) offers particular advantages in terms of complexity and weight reduction. Therefore, it seems very attractive to support the aviation industry in their de‑carbonisation process. Major challenges in terms of operational stability and reliability hinders this technology to be used in commercial applications.Therefore, this study thoroughly examines the conditioning and optimization of open cathode PEMFC stacks, drawing insights from experimental findings. Key areas of focus include the enduring impact of pre-humidification, the quantification of efficiency enhancements through reconditioning via oxygen starvation, and the refinement of Electrochemical Impedance Spectroscopy (EIS) data analysis under non-stationary conditions.The pre-humidification enhances stack performance by improving cell voltages from 40.09 V to 41.37 V at 30 A, increasing membrane humidity and improving efficiency. Furthermore, the residual water in the stack also functions as evaporative cooling and can assist in limiting the operating temperature of the stack during system start up. However, it is demonstrated that excessive soaking with water leads to severe flooding phenomena at the beginning of operation.A reconditioning period of 100 s through oxygen starvation induces a notable increase in stack voltage, from 41.37 V to 47.79 V at 35 A, which decreases to 43.56 V after 10 min of operation. This corresponds to an average increase in electric energy provision of about 6.2% at constant hydrogen consumption, attributed to PtOx reduction and increased water production.Despite limitations outside the medium frequency range (approximately 10 Hz – 15 kHz) for non-stationary conditions, EIS aids in understanding of the stack behaviour and supports the interpretation of current and voltage results. A novel evaluation method enables the quantitative description of the condition using EIS data. This data reveals a considerable drop (on average about 21,5% for the 10 A point and about 27,4% for the 30 A point) in charge transfer resistances of the fuel cell from initial operation to measurements after overnight soaking and the first series of oxygen starvation recovery.

Reliability of Vapor Transport Measurement in a Hollow Fiber Membrane Humidifier

Reliability of Vapor Transport Measurement in a Hollow Fiber Membrane Humidifier

ONE-DIMENSIONAL TRANSIENT MODEL OF FOOD PRODUCT TEMPERATURE AND WEIGHT LOSS EVOLUTION IN A VACUUM COOLER WITH MEMBRANE HUMIDIFICATION

A one-dimensional, transient model was developed to predict the temperature and weight loss evolution of a food product cooled using a vacuum cooling system with membrane humidification. The model included the following: (1) a sub-model based on the mass conservation of air and vapor inside the vacuum chamber; (2) a sub-model based on the coupled mass and heat transfer with inner vapor and heat generation; and (3) a sub-model based on the mass conservation of water vapor along the flow channels inside the humidifier. The vacuum chamber pressure was reduced from 1 atm to 2200 Pa using a vacuum pump with deflating speed of 500 m3/hr and condenser and membrane humidifier operating temperatures of 13°C and 60°C, respectively. A meat joint with a weight of 5.30 kg was cooled from an initial core temperature of 74°C to below 10°C. The conventional and modified models both cooled the food product to 10°C in 25 minutes. It was found out that increasing the chamber relative humidity reduced the surface water evaporation while the internal evaporation rate remained the same. Thus, the meat's moisture content, which was initially at 73.51%, was reduced to only 65.44%-an improvement from the conventional vacuum cooling system's 64.31%. This translated to product weight loss of only 10.98%, which represented significant progress compared to the conventional system's 12.51%. The simulation results indicated that employing humidification in a vacuum cooling system reduces the product total moisture loss while maintaining the rapid cooling rate.

Water Transport and Salt Precipitation in Anion-Exchange Membrane Electrolyzers

The hydrophilic nature of hydrocarbon ionomers and membranes presents a major challenge in anion-exchange membrane (AEM) electrolysis, despite improved stability from reinforcements and modified polymer chemistry. Durability and performance are still highly influenced by the operation mode and water management within the membrane-electrode-assembly (MEA).Dry cathode operation in AEM water electrolysis, where only the anode is supplied with liquid electrolyte (KOH), offers increased hydrogen purity but poses challenges for non-fluorinated anion-exchange polymers that are sensitive to water content. In-situ neutron imaging with high resolution (~6 µm effective resolution) revealed that adjusting the anion-exchange capacity of the cathode binder ionomer can retain membrane humidification (Fig. 1)1.In the electrochemical reduction of CO2 using AEM, water management and salt precipitation are significant challenges. High-resolution neutron imaging of a zero-gap CO2 electrolyzer operating at 200 mA cm-² and 2.8 V showed salt precipitates penetrating the cathode gas diffusion layer, leading to blocked CO2 gas transport and a decay in Faraday efficiency. Salt accumulation was higher under the cathode channel of the flow field than the land2.Both studies demonstrate the substantial impact of water management and salt precipitation on the operation of AEM-based electrolyzers and how neutron imaging can assist in addressing these challenges. References (1) Koch, S.; Disch, J.; Kilian, S. K.; Han, Y.; Metzler, L.; Tengattini, A.; Helfen, L.; Schulz, M.; Breitwieser, M.; Vierrath, S. Water management in anion-exchange membrane water electrolyzers under dry cathode operation. RSC Adv 2022, 12, 20778–20784.(2) Disch, J.; Bohn, L.; Koch, S.; Schulz, M.; Han, Y.; Tengattini, A.; Helfen, L.; Breitwieser, M.; Vierrath, S.; High-resolution neutron imaging of salt precipitation and water transport in zero-gap CO2 electrolysis. Nat Commun 2022, 13, 6099. Figure 1

Noninvasive Local Impedance Determination in Polymer Electrolyte Fuel Cells Stacks

Water management in polymer electrolyte fuel cells (PEFC) needs to ensure a homogeneous membrane humidification without local dryout or flooding from inlet to outlet of cell with active areas of some hundreds of square centimeters. Measurement methods that are able to provide spatial resolved information on the current density, membrane conductivity or water content distributions of PEFCs are usually invasive and therefore difficult to use in actual applications. Based on the learnings of the development of electrical impedance tomography (EIT) for PEFCs [1], a new method is presented that is capable to determine 1D humidity profiles between the gas inlet and outlet in a 6-cell commercial stack [2,3]. Local AC current injections and voltage measurements are performed using electrodes positioned all around the fuel cell attached to the flow field plates (see Figure 1a). Each electrode combination characterizes a specific area of the cell as the current crosses the membrane electrode assembly (MEA) very locally.The presentation will discuss the calibration of the local impedance values, the sensitivity of the approach to relative humidity (rH) gradients in the cell (see Figure 1b) and present use cases of measurements during fuel cell operation. Finally, we will present results on multiply frequency AC stimulation that allows local electrochemical impedance measurements and thereby insights beyond membrane conductivity, i.e. into mass transport limitations due to water accumulations in the cell.

Empirical correlation of water diffusivity in hollow fiber membrane humidifier by measuring latent effectiveness

A hollow fiber membrane serves as a crucial component for the membrane humidifier in proton-exchange membrane fuel cell applications. This study established an empirical correlation of a fundamental characteristic, water diffusivity in the annulus membrane, employing a latent effectiveness-number of transfer units approach under isothermal conditions. Initially, experiments are conducted on a specifically designed membrane humidifier, considering temperature, flow rates, relative humidity, and pressure to determine the latent effectiveness. Subsequently, the overall mass transfer coefficient of a hollow fiber tube bundle, possessing latent effectiveness, is utilized to determine the water diffusivity in the hollow fiber membrane.The empirical correlation of water diffusivity is established through dimensionless parameters, including Reynold and Schmidt numbers. Using measured water diffusivity data, the regression coefficients of the empirical correlation were calibrated. The deviation of the correlation was assessed in terms of the coefficient of determination (R2), which exceeded 0.87, demonstrating a high degree of accuracy upon comparing the correlated diffusivity and experimental data. As substantiated by the results, this correlation provides reliable predictions for an actual membrane humidifier, with a mean deviation of only 7.72% during the validation process. Thus, the present findings will aid the design and operation of membrane humidifiers in fuel cells, improving their performance and efficiency by facilitating precise control of water diffusivity.

Bypass Configurations of Membrane Humidifiers for Water Management in PEM Fuel Cells

Water management is an important criterion in the operation of proton-exchange membrane fuel cells to maintain the high performance and reliability of the system. The water content in the cathode air that is supplied to the cathode channel contributes to the membrane humidification and the transport of protons inside the membrane structure. In automotive applications, the supply air is typically driven through an external membrane humidifier to absorb more moisture from the recirculated cathode exhaust. In the literature, humidifiers and fuel cell stacks have been separately investigated without considering whole-system configurations for water management. This study investigates changes in the cathode air characteristics through a membrane humidifier and compares two configurations using a humidifier bypass of the supply flow and exhaust flow to adjust the cathode inlet air relative humidity. Each component in the system was modeled using mathematical relations and converted into blocks of inputs and outputs in MATLAB/Simulink for simulation. The bypass valve was demonstrated to effectively reduce the relative humidity of the supply air from the saturation rate to above 60%, with a bypass fraction of up to 0.6 in both configurations. These adjustments provide system flexibility to accommodate load changes and prevent flooding in the stack channels. Bypassing the supply air through the humidifier effectively maintained consistent cathode inlet humidity across a wide operational range. A 0.4 bypass fraction on the supply side sustained a relative humidity of around 80% for the whole range of operating flow rates. In contrast, the exhaust-side bypass had a smaller impact, and the relative humidity of the cathode air was reduced when the flow rate and bypass fraction increased. This study further supports the control system design to regulate the bypass fraction according to load transients.

CFD investigation of the effect of flow field channel design based on constriction and enlargement configurations on PEMFC performance

In this study, a three-dimensional CFD, non-isothermal model was developed to identify an optimal layout for key geometrical parameters of PEMFC flow field designs with constriction and enlargement cross-sections located at the middle of the channel. For this purpose, the effect of channel configurations with both constriction and enlargement cross-sections was explored and compared with a conventional constant channel segment, under varying hydraulic diameter and lengths. As results, the constriction configuration channel with a hydraulic diameter reduction of 50% has shown the best performance among the enlargement and straight channels for all considered lengths. The constriction configuration enhances fluid velocity, resulting in improved flow distribution and relatively increased pressure drop across the flow channels. Moreover, a better temperature distribution, higher oxygen consumption, improved membrane humidification, and water removal were observed. This leads to an increased net power output of the PEMFC. However, considering different hydraulic diameters of the constriction geometry, the best performance of the cell was achieved at a hydraulic diameter reduction of 60%, combined with a length of 4/7 of the total channel length. This optimal balance between pressure drops and reactants’ flow has enhanced the net power of the proposed geometry by 26.30 %, compared to the straight channel. The novel proposed single-channel configuration based on a simple constriction, is a low-cost design that has significantly improved cell performance compared to the straight one. Since operating conditions are crucial for cell performance, the effects of operating temperature and relative humidity were investigated, and the optimum conditions were determined.

A general effectiveness-NTU modeling framework for membrane dehumidification systems

Selective membrane dehumidification is a promising technology for energy efficient air dehumidification in buildings and has been ranked as a top alternative technology by the Department of Energy. One critical area that needs to be addressed is system modeling that incorporates device mass transfer characteristics and sizing. This work develops, validates, and applies a simple effectiveness-number of transfer units (ε-NTU) modeling framework for dehumidification technologies that can be used to predict performance and determine required membrane area for a wide range of technologies, including energy recovery ventilators, vacuum membrane dehumidification, and desiccant dehumidifiers. Although the ε-NTU method is a well-established framework for sizing and modeling heat exchangers, past attempts to create εL-NTUL(meaning ε-NTU for latent loads) models for dehumidification (referred to as latent cooling): (1) don’t agree on the definition of NTUL, (2) are not fully analogous to the heat transfer definition of NTU, (3) cannot be easily applied across all technologies, (4) use humidity ratio, not vapor pressure, as the driving force, and (5) are often complex or lack a fundamental basis. In this work, we develop an effectiveness-NTU model for membrane humidity exchangers that is based on vapor pressure differences, employs a novel definition of the latent number of transfer units for dehumidification applications (NTUL) and is fully analogous to the heat transfer effectiveness-NTU models employed for sensible heat exchangers. The resulting εL-NTUL relationships are validated against experimental data with a maximum error of 1.2%. The overall methodology is applied in case studies for vacuum membrane dehumidification system sizing showing that the ratio of required membrane area per building floor area ranges between 0.1 and 3.5 depending on the building type and operating conditions. Finally, the framework is further generalized for applicability to any gas separation application.

Temperature Dependence of Water Crossover in Proton Exchange Membrane Water Electrolysis

The importance of hydrogen production by proton exchange membrane water electrolysis (PEMWE) in the energy sector makes it worthwhile to understand the system behaviour in failure cases and possible recovery procedures. In particular, insufficient feed water supply to the anode can be caused by inhomogeneous distribution in the stack, pump failure or erroneous operation strategy. For the PEMWE, a distinct turnover point has been reported if feed water supply is reduced in isothermal conditions.[1] Further reduction of the water supply results in higher ohmic resistance due to gradual membrane dehydration and increased mass transport losses.[2] Here, we will use the term “water starvation” to describe this event. The drainage of water from the anode can be explained by oxygen evolution reaction (OER), humidification of oxygen gas and water crossover to the cathode. The last one was measured for well hydrated membranes at atmospheric pressure [3] or below typical operating temperatures [4].In the first part of this work, we present the water crossover in a PEMWE on single cell level to understand their dependence on the operating parameters such as temperature and pressure. Media were supplied and conditioned to a 5 cm2 single cell by using an in-house test bench and a potentiostate with booster. The electrochemical measurements were performed with a commercially available catalyst coated perfluorosulfonic acid membranes (loading of 0.3 mg/cm2 geo Pt/C on cathode and 1 mg/ cm2 geo Ir-Oxide on anode). The anode was kept at atmospheric pressure, whereas the cathode pressure was varied from atmospheric to 400 kPaabs at various current densities and cell temperatures of 40 – 80 °C. To determine the water crossover under several operating conditions, the water vapour was condensed to liquid phase in a heat exchanger and measured by a gravimetric method.Furthermore, the measured water crossover is used to force a water starvation in the second part of this work. After cell assembly, a conditioning procedure was performed until steady cell performance was achieved. Cell performance was evaluated using polarization curve, cyclic voltammetry (CV), electrochemical impedance spectroscopy (EIS) and water crossover. The last two are of particular interest, because they provide information about membrane humidification. The Tafel slope and mass transport overvoltage were evaluated from the polarization curves and high frequency resistance (HFR). Water starvation experiments were then carried out at a constant cell voltage by reducing the feed water supply, resulting in a dramatic decay of the current density. The analysis of Tafel slope and EIS data reveals an increase of ohmic and mass transport resistances. Furthermore, the HFR measurements were correlated with water crossover as a function of membrane hydration. We observed that the specific water crossover per current density decreases as a consequence of membrane dehydration. After this step, feed water supply was increased to an unrestricting level for membrane rehydration and performance recovery. For this recovery procedure, different operation modes were employed by varying the current densities. At constant current density, membrane hydration is influenced by the ionic current from the cathode to the anode via proton drag. EIS and water crossover measurements allow us to determine the influence of operation on the dynamics of the membrane rehydration process. Together with the polarization curves, cyclic voltammetry and HFR data, reversible and irreversible degradation processes caused by the water starvation can be identified.Altogether, we present a deeper understanding of water crossover for a wide operating parameter range and provide a guideline for recovery of the performance for PEMWE.[1] Christoph Immerz et al 2020 Meet. Abstr. MA2020-02 2456[2] C. Immerz, B. Bensmann, P. Trinke, M. Suermann, R. Hanke-Rauschenbach, J. Electrochem. Soc. 2018, 165, F1292-F1299.[3] K. Onda, T. Murakami, T. Hikosaka, M. Kobayashi, R. Notu, K. Ito, J. Electrochem. Soc. 2002, 149, A1069.[4] P. Medina, M. Santarelli, International Journal of Hydrogen Energy 2010, 35, 5173.

Distributed Intelligence in Autonomous PEM Fuel Cell Control

A combination of perceptive and deliberative processes is necessary to ensure the efficient and autonomous control of proton exchange membrane fuel cells (PEMFCs) under optimal humidification conditions. These processes enable monitoring and control tasks across various application scenarios and operating conditions. Consequently, it becomes crucial to adjust parameter values corresponding to different states of the PEMFC during its operation. In this context, this work presents the design and development of an architecture for the control and management of a PEMFC with a maximum power output of 500 [W] based on intelligent agents operating under optimal conditions (membrane humidification). The proposed architecture integrates perception and action algorithms that leverage sensory and contextual information using heuristic algorithms. It adopts a hierarchical structure with distinct layers, each featuring varying time windows and levels of abstraction. Notably, this architecture demonstrates its effectiveness in achieving the desired energy efficiency objective, as evidenced by successful validation tests conducted with different electrical power values delivered by the fuel cell, encompassing three distinct operating states (dry, normal, and flooded). An exemplary application of this scheme is the dynamic control of the humidification of the polymeric membrane, which further highlights the capabilities of this architecture.