This thesis performs a sensitivity analysis and an error propagation for a simulation model of the LSS on the ISS, which has been developed in the simulation environment of V-HAB. V-HAB results from a long year research work at the LRT of the TUM. The general task is the examination of variations of certain simulation parameters, whereby the possible error of the simulation compared to the real system should be estimated. If possible, an error propagation calculation should be done thereby.
Therefore, in a first step it is necessary to identify the simulation parameters, which allow a variation because of not exactly known technical specifications, natural oscillations or certain modelling assumptions, and to modify the simulation so that a variation of the parameters is possible and comprehensible. The feasibility of the new implemented simulation values is tested with test simulations and the impact on the simulation results is proven by a sensitivity analysis for plausibility and accuracy.
Already known techniques for error propagation are investigated and the best fitted one for the existing problem is chosen and adapted. Hence a method for the creation of functions to calculate deviations because of variations in simulation parameters is developed and a short guidance for this is provided (Standard approach). Furthermore, this method is modified, to get another calculation approach (Advanced approach), which promises better results. Moreover, an approach to determine errors of the subsystem models and their classification during the error calculation is exemplary presented.
As an example, these calculating methods are performed for the deviation of the relative humidity and deviation-functions are generated.
For verification of the developed functions some comparative simulations are done, which are compared with the results from the error calculations and so the accuracy of these functions is estimated.
The Standard Approach yielded mostly poor results during these comparison simulations and the good results were usually random and not reliable. With the Advanced Approach, better and above all more comprehensible results could be achieved. For the estimating of deviations, which were caused by single changed parameters, very accurate values could be calculated. The results with several changed parameters were sufficient to determine the magnitude and direction of the deviations. In the end, an area where the results of the simulation vary due to the changes, could be specified. This also resulted in an estimation of the deviation of the simulation results from reality, which still depends on unknown error values, such as the subsystem errors.
«
This thesis performs a sensitivity analysis and an error propagation for a simulation model of the LSS on the ISS, which has been developed in the simulation environment of V-HAB. V-HAB results from a long year research work at the LRT of the TUM. The general task is the examination of variations of certain simulation parameters, whereby the possible error of the simulation compared to the real system should be estimated. If possible, an error propagation calculation should be done thereby.
The...
»