Gas‐Phase Velocity Estimation in a Tube Bundle by means of Wire‐Mesh Sensors
Tutor / director / evaluatorPrasser, Horst‐Michael
Document typeMaster thesis (pre-Bologna period)
Rights accessRestricted access - author's decision
Understanding of flow structure and hydrodynamics in tube bundles is crucial to develop and safely operate Steam Generators in Nuclear Power Plants. During normal operation, the Steam Generator contains a two‐phase water‐steam mixture at high pressure and saturated temperature (namely 60atm and 275°C). However, in case of tube rupture, completely new conditions arise. The TRISTAN project aims to investigate a simplified model of this scenario using a twophase air‐water flow at ambient pressure and slightly elevated temperature. A Steam Generator mock‐up facility has been already built and commissioned, performing tests with pool scrubbing setup (i.e. a single tube instead of a bundle). The flow analysis method used in TRISTAN is the wire‐mesh sensor, an intrusive flow imaging technique which measures the local instantaneous conductivity of a fluid. Provided each phase of the flow has very distinctive conductivity (e.g. air and water), the proportion of each can be measured with high temporal and spatial resolution. Wire‐mesh sensors deliver three‐dimensional arrays of void fraction which can be used to characterize two‐phase flows. However, WMS do not measure gas‐phase velocity and therefore the z‐axis remains a temporal axis, i.e. bubble size cannot be completely calculated without further assumptions about the velocity. A common approach to solve this problem is the use of two independent WMS, delivering two consecutive sets of void fraction arrays. These arrays can then be cross‐correlated in order to find time‐shifted patterns corresponding to the gas‐phase velocity. To that end, several experiments have been performed with different boundary conditions (namely break type, flow rate and break – sensor distance). Some of these experiments are analysed with the developed velocity estimation algorithms, testing coherence in results and range of applicability. The methods presented in this thesis can be used to compute local instantaneous gas‐phase velocity. This velocity leads to complete bubble identification (i.e. the temporal z‐axis can be transformed into an Eulerian axis), so flow structure and hydrodynamics can be analysed in detail through e.g. bubble size distribution or interfacial area concentration.