One of the major safety issues of Swedish nuclear power plants is the ageing process of the equipment accelerated by an occurrence of thermal fatigue. This is particularly important in view of the planned upgrades of nuclear reactors to higher power levels. To address this issue, it is proposed to perform a development of numerical methods to predict the thermal mixing and fatigue phenomenon. The proposed project will help to develop CFD models to predict amplitudes and frequencies of temperature fluctuations in mixing turbulent flows. For this purpose, a large eddy simulation, inclusive of conjugate heat transfer between fluid and solid domains, has already started. However, larger computational resources are required to estimate the aforesaid temperature fluctuations. The computational results will be validated against experiments run at KTH.