SIMULATION OF THE SELF-SUSTAINED OSCILLATIONS OF FLOW AND HEAT TRANSFER IN A PERIODIC CORRUGATED CHANNEL USING TURBULENT MODELS

摘要

Within the periodic channel, the turbulence characteristic of the flow and heat transfer is obvious due to the strong perturbations caused by the geometry of the channel at a very low Reynolds number（Re≥600）. According to the previous studies, the unsteady Laminar model was used to simulate the flow and heat transfer in the periodic channel to research the self-sustained oscillations. On this basis the author also studied the unsteady character using a direct numerical simulation when the Reynolds numbers range from 100 to 20000. The results showed that the flow and heat transfer was stable at a low Reynolds number, and it became unstable with self-sustained oscillation when the Reynolds exceeded to a threshold value. In this paper, the turbulent models were used to simulate the flow and heat transfer in the periodic channel Based on the result of the previous studies. Different turbulent models were used at the same Reynolds number with the same time-step during the simulation in this paper to discuss the unsteady character of the turbulent flow. It is indicated by compare that the results of simulation can present the property of unsteady turbulence when the time step size is less than 1e-3. The velocity U varies with time. However in consideration of flow field, it was similar with the result of direct simulation when the time step △t≤1×10-5.while because of the time-average, it is different when the time step was larger than 1×10-5. The turbulent character of the flow and heat transfer can not be obvious. During the Reynolds-averaging equations simulation, four different turbulent models were used: Spalart Allmaras models, standard k-ε model, RNG k-ε model and also Realizable k-ε model. It was indicated that the time-average of the unsteady flow obtained in the computation still possess the property of self-sustained oscillations. Considered that the physical quantities were handled in time-averaging method in the turbulent models was much better to reflect the distribution of the flow and the temperature field. The variation of Nu with Reynolds numbers and U with time was consistent with the result of direct simulation. That is the U of a certain spot in the channel varied with time at a certain amplitude and frequency which proved that the flow was unsteady and the time-average also had self-sustained oscillations.