Today, there is an indispensable need in alternatives to fossil energy sources because its availability is timely limited. In addition to the sooner or later arising availability shortage, environmental effects caused by the utilization of fossil fuels are essential driving forces in pursuing a sustainable alternative. In this context, hydrogen is often mentioned as one of the most promising energy carriers of the near future and it is seen as potential emission-free fuel in the transportation sector. Unfortunately, most common hydrogen production methods are either based on fossil fuels or the energy intensive water electrolysis. These methods cannot fulfill modern energy provision requirements, because they are neither sustainable nor energy efficient. One possibility of a sustainable and energy efficient hydrogen production is the biotechnological pathway of dark fermentation, where biomass can be used to generate a hydrogen-rich gas. Besides hydrogen as main component carbon dioxide is also generated with its ratio depending on the respective process conditions. The carbon dioxide in the gas is considered as a contaminant and has to be separated subsequently to obtain pure hydrogen as a final product. Typically used hydrogen purification methods are amine absorption and pressure swing adsorption. These large-scale industry methods are well-established and may be suitable from a technological point of view. But, with gas capacities, compositions, and conditions different from typical applications, these methods might not be suitable for the purification of dark fermentation gas regarding their energy efficiency. For this particular application membrane technology seems to be very promising because gas permeation is a simple, flexible, and even mobile purification method. It is also characterized by a low energy requirement and low investment costs. Therefore, the goal of this work was to investigate the applicability of gas permeation for the purification of a dark fermentation gas. An innovative small scale process to upgrade the hydrogen rich gas was designed, with special attention to a flexible and energy-efficient separation process setup. Two areas of research were covered: an experimental investigation of commercially available membrane material, and a simulative investigation of different process setups. The experimental part included assembly and testing of membrane modules in the laboratory, using either PI or PP hollow fibers as membrane material. Pure gas measurements for PI membrane modules resulted in ideal H2/CO2 selectivities in the range of 3.45 to 3.91. Compared to that, using PP membrane material never surpassed an ideal H2/CO2 selectivity of 2.7, which made them unsuitable for the given purification task. Further investigation of PI membrane modules included measurements with two different mixtures, a binary with 66 vol% H2 and 34 vol% CO2, and a ternary with 40 vol% H2, 30 vol% CO2 and 30 vol% N2. These investigations showed that for binary or ternary gas mixtures the respective H2/CO2 selectivities decreased to values in the range of 2.29 - 2.49 or 2.43 - 2.7, respectively. The laboratory tests were followed by the design of a separation process to set up a small scale pilot plant. This pilot plant was then connected to a hydrogen fermenter and the obtained information on process data and gas qualities confirmed the results from laboratory measurements. Based on the experimental work it can be concluded that an actual online fermenter gas upgrading is possible. In the simulative part of this work, the simulation tool Aspen Custom Modeler® was used to develop a gas permeation unit operation which was implemented into the flow sheeting and process optimization software Aspen Plus® at a later stage. This single-stage model was successfully validated and used to design various multi-stage processes in Aspen Plus®® . Implementing a second (setups 1 and 2) and third stage (setup 4) into the process resulted in a H2-recovery increase compared to a single-stage configuration. The models were able to display the behavior of the membranes due to changing process conditions, but when using H2-selective membranes the achievable product quality was unfortunately not reached. Compared to setups 1 and 2, the utilization of CO2-selective membranes (setup 3) did reach the required product purity of at least 98 vol%. It furthermore resulted in a lower specific energy demand. All in all, the purification of fermentatively produced hydrogen to a hydrogen combustion engine suitable gas quality could only be maintained with CO2-selective material. Possible scenarios for the off-gas utilization were developed, to gain insight into the available unused energy potential. For all four multi-stage setups, it was shown that a thermal utilization of the off-gas- energy content could provide a significant amount of heat and power. In conclusion, this thesis showed that the utilization of commercially available membrane material has its possibilities but also its restrictions. Regarding sustainability and emission reduction, this pathway of fermentative hydrogen with subsequent membrane purification seems very promising.