Abstract: With the relentless pursuit of higher thrust to weight ratios and superior operational reliability in modern aero engine development, the turbine disk, being one of the most critical rotating components of the engine, has drawn ever increasing research attention. Given its severe service environment involving complex thermomechanical coupling effects, the structural design of the turbine disk light weight and high performance is of paramount importance for ensuring overall engine efficiency, durability, and safety. In the density-based topology optimization, it is challenging to impose boundary dependent loads, such as convection, due to the lack of explicit boundary representation. To solve this issue, this paper proposes a method for identifying internal and external boundaries of turbine disks and imposing convection through the advection diffusion equation and density gradients. By solving the advection diffusion equation under lighting on both sides of the disk, the obtained shadow regions are used to indicate both internal cavities and external regions of the structure. Subsequently, the internal and external boundaries are identified in conjunction with density gradients. Through the use of density gradient information, the convective heat transfer loads on the boundaries are transformed into body-wise heat source terms and imposed in an implicit manner, which completely eliminates the requirement for explicit boundary tracking. In the optimization modeling, a global stress constraint that accounts for design dependent loads including thermal convection, thermomechanical coupling, and centrifugal force is constructed via Heaviside function projection, and the adjoint method is employed for sensitivity analysis. Based on the established topology optimization model for the axisymmetric thermomechanical coupling problem, numerical solutions are obtained for a typical turbine disk configuration. The effects of varying convective boundary conditions on the optimal topological layouts and stress distributions are analyzed in detail, and cross validation of the optimization results is performed. Numerical results demonstrate that it is imperative to consider thermal convection in turbine disk topology optimization, and the proposed boundary dependent thermal convection loading method proves to be effective and feasible.