Abstract The influence of g-jitter onboard the International Space Station (ISS), on diffusion in liquids and more specifically on thermodiffusion experiments has been a controversial issue. Meanwhile, only a few studies have implemented real acceleration data recorded during a specific experiment. This study, numerically surveys a typical thermodiffusion experiment that has been performed on ISS in late 2009 (IVIDIL-Run2). The simulations employed the real acceleration data recorded in the same time for both residual (quasi-steady) and g-jitter accelerations. Different interval methods have been applied to extract the data from g-jitter for the fluid dynamic simulations. It was obtained that the results of interval Root Mean Square (RMS) is not realistic. Furthermore, the current study suggests a relation with the low-frequency contribution of the g-jitter and the deviations in the separation process throughout the experiment. It was obtained that low-frequency contribution of g-jitter is most-likely responsible for disturbing ideal diffusion process.