In turbomachinery design, the accurate prediction of the life cycle is one of the most challenging issues. Traditionally, life cycle calculations for radial turbine wheels of turbochargers focus on mechanical loads such as centrifugal and vibration forces. Due to the increase of exhaust gas temperatures in the last years, thermomechanical fatigue in the turbine wheel came more into focus. In order to account for the thermally induced stresses in the turbine wheel as a part of the standard design process, a fast method is required for predicting metal temperatures. In order to develop a suitable method, the mechanisms that cause the thermal stresses have to be understood. Thus, in a first step, a detailed analysis of the temperature fields is conducted in the present paper. Extensive numerical simulations of a thermal shock process are carried out and validated by experimental data from a test rig. Based on the results, the main heat transfer mechanisms are identified, which are crucial for the critical thermal stresses in transient operation. It is shown that these critical stresses mainly depend on local 3D flow structures. With this understanding, two fast methods to calculate the transient temperatures in a radial turbine were developed. The first method is based on a standard method for transient fluid/solid heat transfer. In this standard method, heat transfer coefficients are derived from steady-state computational fluid dynamics (CFD)/conjugate heat transfer (CHT) calculations and are linearly interpolated over the duration of the transient heating or cooling process. In the new method, this interpolation procedure was modified to achieve an exponential behavior of the heat transfer coefficients over the transient process in order to enable a sufficient accuracy. Additionally, a second method was developed. In this method, the specific heat capacity of the solid state is reduced by a “speed up factor” to shorten the duration of the transient heating or cooling process. With the shortened processes, the computing times can be reduced significantly. After the calculations, the resulting times are transferred into realistic heating or cooling times by multiplying them with the speed up factor. The results of both methods are evaluated against experimental data and against the results of a numerical method known from literature. The methods show a good agreement with those data.
Read full abstract