Identification of flow regime using multi-resource information

Most two-phase flow measurements depend on correct flow regime identification. Many methods have been proposed to identify the flow regime, but none of them can work well across different flow conditions. In this paper we apply a non-intrusive instrument of different types of sensor to extract the data under different flow conditions. In light of these different sensor information, evidence theory of Dampster-Shafer is applied to realize the information fusion. Based on a clustering technique of self-organized feature map, the basic probability assignment for Dampster-Shafer is determined, and at the same time the calculation of combination rule using the evidence theory can result in new interpretation and insight of different flow regimes. The experimental results show that the principle of data fusion helps improve the identification quality of flow regimes greatly. Key-Words: flow regime; two-phase flow; evidence theory.