Effect of injector geometry on two phase flows is of profound importance to industry. If the injection method is found to vary the flow characteristics dramatically, it can be employed to obtain desirable two phase flow regimes/attributes and avoid rather unsought conditions. This could potentially save a lot of costs in the extraction and transportation of oil and gas as well as in many other applications. Moreover, the issue of flow development and dependency on the injection conditions is essential when modelling two phase flows. A lot of experimental data and empirical models may have been developed based on systems that may not be fully developed. Therefore, inaccurate modelling of the physical interactions of the flow gets adopted, and hence large divergence between models and experimental data produced by different researchers often transpires.
Most of the published studies on entrance effect were conducted on air-water or steam-water systems because of their relevance to heat transfer units in the nuclear industry. This paper presents an extensive experimental investigation into the issue of flow development using two approaches; measuring void fraction at five axial stations along the test section, and using different geometries for bubble injection into the base of the pipe. The study focuses on how the entrance effect is influenced by the liquid viscosity. The experiments were conducted in a 127 mm diameter vertical pipe. The investigation is achieved by contrasting 180 runs produced using three different injector geometries, the runs were repeated using 4 different oil viscosities, making 2160 experimental run. Gas superficial velocity (Ugs) was varied between 0.01–5.40 m/s, while liquid superficial velocity (Uls) between 0.07–0.86 m/s. The viscosities investigated span between 4.0 cP up to 104.6 cP. The void fraction was measured using Electrical Capacitance Tomography (ECT) and the Wire Mesh Sensor (WMS). That in addition to differential pressure measurements.
Validation of NEPTUNE-CFD Two-Phase Flow Models Using Experimental Data
