**5. Conclusion**

To perform thermal hydraulic design of the boiling water reactor (BWR) without actual size tests, we have been developed a new design method for BWRs. In this design method, the two-phase flow correlations for fluid mixing phenomena must be modified or created based on results of large scale numerical simulations. Then, we developed an innovative two-phase flow simulation code TPFIT and an advanced interface tracking method. In the advanced interface tracking method, gas and liquid mass conservation equations are solved, to treat compressibility of gas and to keep high volume conservation of gas and liquid.

In the second step, to verify and validate the TPFIT with the advanced interface tracking method, the TPFIT was applied to some experimental analyses including three the 2-channel fluid mixing tests, and comparisons between measured and predicted result were carried out. By these results, it was concluded that the TPFIT can be applied to the two-phase flow fluid mixing phenomena.

In the next step, the TPFIT was applied to the numerical simulation of fluid mixing between subchannels on a current BWR and FLWR that is one of innovative water reactor concept developed in JAEA. The existing two-phase flow correlation for fluid mixing was evaluated using detailed numerical simulation data. When inlet quality ratio of subchannels is relatively large, evaluated mixing coefficients by existing two-phase flow correlations for fluid mixing were different from those of the detailed numerical simulation data. And new correlation based on large scale numerical simulation results are expected.

In the fourth step, we tried to develop new two-phase flow correlations related to the fluid mixing phenomena. Differential pressure fluctuation between subchannels dominates the fluid mixing phenomena. We considered that the differential pressure fluctuation is induced by difference of axial pressure distribution in each subchannel. Then we developed new correlation for axial pressure distribution. In this correlation, a hydrostatic gradient, acceleration and frictional gradients are taken into account in the liquid slug, and no pressure gradient exists in the slug bubble. New correlation was validated by numerical results of the TPFIT, and it was concluded that predicted results of this correlation almost agree with the simulation results. Furthermore, differential pressure fluctuation based on predicted axial pressure distribution in subchannels can predict differential pressure by the TPFIT. We thought that the method to create new twophase flow correlations by results of large scale numerical simulations is effective. However, there is issue such as required computational resources or validation of prediction accuracy of created correlations.
