| Popis: |
Obstoječi matematični modeli kapilarnih cevi se pri določanju koeficienta trenja zanašajo na splošne korelacije, ki niso bile zasnovane za tok hladiva skozi te cevi. Obenem ni jasno, ali je dvofazni tok v njih homo oz. heterogen. V tem delu je zato predstavljen razvoj izboljšane korelacije, ki je bila vključena v numerični model kapilarne cevi. Točnost slednjega se je s tem v povprečju povečala za ±1,3 %. Narejena je bila tudi vizualizacija toka, ki je potrdila, da je ta heterogen čepast tok. V nadaljevanju je bil zasnovan še algoritem za optimizacijo diabatnih kapilarnih zvez. Primerjava različnih optimizacijskih algoritmov je pokazala, da so najbolj zanesljivo delovali t. i. evolucijski optimizacijski algoritmi. Existing capillary tube models rely on general friction factor correlations, which were not designed for the flow of refrigerant through these tubes. Additionally, it is unknown whether the two-phase flow in them is homo- or heterogeneous. This work presents the development of an improved friction factor correlation, which was implemented in a numerical capillary tube model. The accuracy of the latter increased by ±1,3 % as a result of this. A visualization of the flow was also performed. It showed that the flow is a predominantly heterogeneous plug flow. In addition to the performed experiments, an algorithm for the optimization of capillary tube-suction line heat exchangers was also developed. Comparison of different optimization algorithms showed that evolutionary algorithms performed most favourably. |
