Study On Turbulence Model Under Supercritical Pressure

Supercritical water reactor is one of the Generation IV reactors, which is regarded as an innovative system due to its high thermal efficiency and simplified direct steam cycle. The thermal-hydraulics of supercritical fluid is an important topic of supercritical reactor's design.Heat transfer at supercritical pressure is mainly characterized by the thermal physical properties which vary strongly, especially near the pseudo-critical line. According to different heat transfer mechanisms, there are several types of heat transfer. First, Heat transfer at high mass flow rate and low heat flux, which is regarded as forced convection. It is found an enhancement in heat transfer. Second, Heat transfer at low mass flow rate and high heat flux, which is regarded as mixed convection. Buoyancy influences heat transfer greatly. Third, Heat transfer at high mass flow rate and heat flux. The heat transfer ability is impaired by acceleration.In this article, forced and mixed convection with heat transfer in vertical tube under supercritical pressure is studied with numerical simulation. The main work of numerical study is assessment and improvement of turbulence models, because fluid under supercritical pressure is single phase. Low Reynolds number k-εmodel has good application prospects. Several low Reynolds number k-εmodel is assessed, and according to buoyancy mechanism, turbulence model is improved in this article.For forced convection, the results of computation indicate that traditional Low Reynolds k-εmodel can predict wall temperature correctly. It is considered that heat transfer is determined by Prandtl number near the wall based on numerical result. When the Prandtl number becomes maximums, the temperature viscous layer becomes thinnest. It leads to heat transfer enhancement.There are two types of mixed convection, buoyancy aided-flow and buoyancy opposed-flow. For buoyancy aided-flow, heat transfer ability is impaired, recovered and impaired again under buoyancy effect. The computation results show that the traditional turbulence models can qualitatively simulate the trend of wall temperature, but cannot predict wall temperature correctly. DNS data show that the buoyancy term in k equation is dominant. Traditional buoyancy term model underestimates buoyancy term greatly, so it cannot predicted heat transfer quantitatively. In this article, an improved AFM is adapted as buoyancy model, and the turbulence model is improved. A new k-ε-kt-εt model is proposed. The results show that present model performs much better than traditional models. It is found that buoyancy effect modify shear stress and velocity gradient. This effect is called external effect, which means buoyancy indirectly influence turbulence. When the shear stress is suppressed and velocity gradient is decreased under buoyancy, the heat transfer impairs. When shear stress increases and velocity distribution becomes M-shape, the heat transfer recovers. Also, buoyancy directly influences turbulence, which is called structural effect. In k equation, buoyancy term is an important term, which means the buoyancy effect is a main sauce of turbulent kinetic energy.For buoyancy opposed-flow, buoyancy effect enhances heat transfer. Theory analysis and DNS data show that the buoyancy term in k equation is negligible. So the choose of buoyancy term model is not very important. Traditional model can provide right wall temperature. It is found that buoyancy effect leads to increasing shear stress and increasing velocity gradient, these effect enhance heat transfer. It means external effect is very important. And the structural effect is negligible.