This thesis deals with the numerical evaluation of flow inside the MBU-20/P aviation oxygen mask, with a focus on the amount of carbon dioxide re-inhaled by the user. The theoretical part describes the principles of aviation oxygen masks, the issue of mechanical dead space, CO₂ rebreathing, and the influence of internal flow on the washout of the mask cavity. The practical part focuses on the digital reconstruction of the mask geometry, preparation of the computational domain, and transient CFD simulation of representative breathing cycles.
The monitored quantity was the time-dependent CO₂ concentration in the mouth region during the inhalation phase. For quantitative evaluation, the area under the CO₂ concentration curve during the inhalation interval was used. This value was then used as the objective function for optimizing the valve position in the UPTIMAI environment. The design variables were the valve displacements in the X, Y, and Z directions.
Based on the CFD simulation, a surrogate model was created, allowing the influence of valve position to be evaluated rapidly without the need for repeated calculations in Fluent. The accuracy of the surrogate model was verified using R² and RMSE values and subsequently by a control calculation of the area under the curve from data obtained in Fluent. The results showed that the valve displacement in the Z direction had the most significant influence on the value, while the influence of the Y direction was minimal.
The optimized valve position led to a reduction in the area under the CO₂ concentration curve during the inhalation phase compared with the average original curve obtained from 30 models. The results confirm that the valve position can significantly influence the amount of CO₂ remaining in the mask and subsequently re-inhaled by the user. This thesis therefore presents a verified workflow combining digital reconstruction, CFD simulation, and surrogate modeling for the evaluation and optimization of flow in an aviation oxygen mask.