Pipelines are widely used in energy power system including thermal power plants and nuclear power plants. In the power system, especially nuclear power plants, as thermal stratification characteristic of nonuniform temperature distribution in the pipe and thermal fatigue such as the low cycle thermal fatigue due to existence of the cycle thermal stress are inevitable, the pipe line system can be destroyed easily and thus affect the normal operation of the power plants. In order to study the pipeline thermal fatigue, the pipeline thermal stress needs to be calculated and therefore the temperature distributions especially the inner wall temperature was needed. The outer wall temperature and working fluid temperature can be obtained with installing measuring tools. The key and difficult point is the estimation of the inner temperature distribution of the pipe. At the same time, in the practical engineering, the pipeline structure with special safety requirements or higher requirements for structural completeness are not allowed to be destroyed by the measuring equipment. Based on the consideration above, this paper presents a method of solving the inverse heat conduction problem, which means the inner temperature distributions can be derived by the outer wall temperature distributions. For the pipe inverse heat conduction problem, this paper applies numerical analysis as the main way. Firstly, the method for transient inverse heat conduction problems applying separation of variables and Duhamel’s theorem is established. As the effects of the random error on the measured outer wall temperature are inevitable, the measured data need to be smoothed before used as an input. The Gram orthogonal polynomial method based on the digital filtering theory is applied in this paper to accomplish the smoothing process. The inverse process is accomplished by using MATLAB programming. Then this method is verified in the experiment with high temperature and high pressure. In order to directly validate the accuracy of the inverse analysis, for the test section, not only the transient outer wall temperature and fluid temperature were measured, but also the time dependent middle layer temperature were measured. Then the middle layer temperature obtained from inverse calculation was compared with the measured data from the experiment. The calculated results show that the accuracy of this method is high. The temperature distributions along the radical direction can be obtained quickly and accurate instantaneous heat load for the structural stress analysis and thermal fatigue analysis can be provided using this method. (CSPE)
An investigation of wall temperature characteristics to evaluate thermal fatigue at a T-junction pipe