FUZZY LOGIC INTERFACIAL TRANSFER IN TWO-PHASE FLOW
For the accurate analysis of the two-phase flow, proper models of interfacial transfer between two phases are crucial. However, they are strongly dependent on the flow regimes so that some spurious numerical instability has been experienced on the boarder of the flow regime. In order to relax such an abrupt change of the models due to the flow regime transition, the present study was performed to generate the fuzzy membership functional representation of the flow regime map of the two-phase flow. For the objective approach, the neural networks are employed. The excitation of the output nodes which are correspondent with the representative flow regimes are used in the construction of the membership function. It was found that smoother transition of the interfacial transfer terms are produced by the fuzzy logic reasoning based on the membership functions of the both flow regimes near the transition. Therefore, the present fuzzy representation of the flow regime map can be recommended for the computer codes to enhance their stability of the convergence and accuracy.