Improving the estimation of slug frequency from two-phase flow simulations in horizontal pipelines

Abstract Slug flow is a common flow pattern that is often accompanied by undesired effects, such as important pressure loss or vibrations, leading to large errors in multiphase flow metering. As these undesirable effects strongly correlate with the frequency of slug occurrence, this parameter is of particular interest. In this paper, different methods for calculating the slug frequency are applied to data from multiphase flow simulations and corresponding high-speed video observations. Commonly used methods, such as power spectral density or calculating the mean slug frequency by counting slugs, are compared with new evaluation methods proposed by the authors. These new methods are based on the distribution of the time difference between consecutive slugs. Since these methods place more weight on frequencies that occur more often, they are less sensitive to outliers in the slug frequency distribution. For the developed slug flow, good agreement is observed between the different slug frequency calculation methods and between the experimental data and simulation results. In contrast, significant differences between simulations and experiments can occur when the flow pattern is still developing. In this case, the deviations in the resulting slug frequencies when different calculation methods are applied can indicate how much one can trust the results. Comparing the results derived by the different methods with reference values, which were obtained by visually counting the slugs in the videos, shows the best agreement for the newly introduced peak slug frequency, with a relative error of less than 10% on average and circa 24% maximum.