Thermal stratification is a common and natural phenomenon in energy storage tanks. This paper presents a Computational Fluid Dynamics (CFD) analysis of thermal stratification within a novel cold energy storage, termed a Thermal Diode Tank (TDT). The TDT is a well-insulated water tank equipped with heat pipes for unidirectional heat transfer from water to the ambient air. When integrated into a Refrigeration and Air-conditioning (RAC) system, it provides cooling water to reduce the condensing temperature. The research problem revolves around the impact of thermal stratification on TDT performance and the identification of the optimal TDT structure when thermal stratification is present. Therefore, this study aims to employ CFD to simulate temperature profiles within a TDT, both with and without an obstacle. With validated models, parametric analysis has been conducted to determine the optimal TDT structure. Several design parameters have been investigated, including tank orientation, heat transfer capacity, the ratio of obstacle axial height to TDT height (ho/H), and the ratio of obstacle hole diameter to TDT diameter (do/D). Key findings reveal that a vertical tank orientation is more conducive to improving the TDT assisted RAC system performance than a horizontal one. Increasing the TDT's heat transfer capacity also positively contributes to the system performance. Parameters of an optimal TDT structure include a ho/H ratio of approximately 0.43 and a do/D ratio of 0.4. It is found that the do/D ratio should be at least 0.15 to obtain positive stratification. Furthermore, the presence of thermal stratification can increase the system's Coefficient of Performance (COP) by up to 10 %.