Gas–liquid separators are one type of surface facility among those used in oil fields. In this paper, the study of gas–liquid separation in a cylindrical cyclone separator (GLCC) using computational fluids dynamics was carried out. The multiphase mixture model and the k–ε turbulence model in an air–water mixture with different geometries of the separator varying the inlet angle from 27° to 36° and 45° were used. Later, variables for the volumetric fraction, velocity, and pressure drop in the separator were studied. Finally, a natural gas mixture from a Colombian oil field was simulated using a species transport model. The results showed that a 36° inlet is the most suitable for the separation process due to its capacity to form a high-intensity swirl without produced liquid carry over. Also, it was found that the centrifugal separator could be a suitable alternative compared with conventional gravitational gas–liquid separators.

Tsunami generation and propagation mechanisms need to be clearly understood in order to inform predictive models and improve coastal community preparedness. Physical experiments, supported by mathematical models, can potentially provide valuable input data for standard predictive models of tsunami generation and propagation. A unique experimental set-up has been developed to reproduce a coupled-source tsunami generation mechanism: a two-dimensional underwater fault rupture followed by a submarine landslide. The test rig was located in a 20 m flume in the COAST laboratory at the University of Plymouth. The aim of the experiments is to provide quality data for developing a parametrisation of the initial conditions for tsunami generation processes which are triggered by a dual-source. During the test programme, the water depth and the landslide density were varied. The position of the landslide model was tracked and the free surface elevation of the water body was measured. Hence the generated wave characteristics were determined. For a coupled-source scenario, the generated wave is crest led, followed by a trough of smaller amplitude decreasing steadily as it propagates along the flume. The crest amplitude was shown to be influenced by the fault rupture displacement scale, whereas the trough was influenced by the landslide’s relative density.