对内圆筒高温且旋转的Taylor-Couette系统内冷水的对流传热进行直接数值模拟，旨在探究密度倒置特性的影响。选用多个密度倒置参数，研究从自然对流到湍流的流态演化。结果表明，随着雷诺数的增大，流动经历了2次流态转变：第1次是从浮升力主导的大尺度环流转变为螺旋涡流，密度倒置效应越显著，轴向流动对径向热传输的抑制作用越强，在{\Theta }_{\mathrm{m}}=\mathrm{0.5}附近存在传热极小值；第2次流态转变发生在离心力主导的区域，流动由波状涡流发展为湍流波状涡，产生Görtler涡，有效增强了传热。从密度倒置效应、离心效应和浮力效应联合作用的角度探讨流态转变的物理机理。结果表明，密度倒置效应越强，浮力效应越弱，流动越易在较低雷诺数下提前进入离心力主导的流态。

This paper conducts a direct numerical simulation of the convective heat transfer of cold water in a Taylor-Couette system with a high-temperature and rotating inner cylinder， aiming to elucidate the flow features and heat transfer characteristics under different density inversion parameters. For this purpose， multiple density inversion parameters were selected to study the evolution of flow states from natural convection to turbulence. By examining the hydrodynamic and thermal behaviors of cold water across various parameter variations， this paper reveals the diversity of flow regimes and the tight association between heat transfer characteristics and flow features. With the increase in Reynolds number， the flow undergoes two transitions. The initial transition is from a buoyancy-dominated regime to a spiral vortex flow. Pronounced density inversion effects intensify the axial flow’s suppression of radial heat transfer， manifesting a minimal value of heat transfer near {\Theta }_{\mathrm{m}}=\mathrm{0.5}. The subsequent transition occurs within the rotation-dominated regime， where the flow progresses from wavy vortex flow to wavy turbulent vortex， accompanied by the emergence of Görtler vortices and an increase in smaller-scale structures at elevated Reynolds numbers， thus enhancing heat transfer capability. The mechanisms behind flow regime transitions are explored from the perspective of the combined effects of density inversion， centrifugal force， and buoyancy. The results suggest that a marked density inversion effect weakens the buoyancy effect and strengthens the sensitivity to the centrifugal force， leading to an earlier onset of centrifugal-dominated flow at a lower Reynolds number.