Identification of error sources in fatigue analyses for thermal loadings

To identify thermal loadings (thermal shocks and thermal stratification), in German NPPs, special fatigue monitoring systems have been installed. The detailed temperature measurement uses sheathed thermocouples, which are located on the external component surface. Tightening straps are used for the widespread method of locking the thermocouples into position. The calculation of material fatigue for a loading sequence has to be carried out based on the measured temperature profile of the outer component surface. Should the analysis comply with the ASME III code, Section NB, alternatively the Articles NB-3200 or NB-3600 can be applied. In fatigue analyses based on the outer-surface temperature, the thermal situation at the inner-surface has to be determined (inverse temperature-field calculation). This leading analysis step is not regulated in the ASME III code. Using general purpose finite element programs, this problem cannot be explicitly solved, because it requires knowledge of the thermal situation at all boundaries (temperature or heat transfer). In the frequently practiced method in a finite element calculation, the inner surface temperature profile is varied until a satisfactory compliance of the calculated outer surface temperature with the measured profile is obtained. Since the input parameters are derived from a variable field, the variation process is large-scale and non-explicit (another input-configuration may cause a similar outer surface temperature). Furthermore, the remaining deviation cannot be quantified regarding the resulting error in the calculated material fatigue. Five typical thermocouple installation methods existing in German LWRs were compared and evaluated regarding the quality of outer surface temperature acquisition. With the evaluation of the experimental data, the essential finding is that for the test transients the maximum of the true outer surface temperature change rate is registered incorrectly with all thermocouple installations. In this context, it has to be pointed out that the change rate registration degrees are valid exclusively for the specific test conditions, i.e. for very fast medium temperature changes (thermal shocks) in a pipe with the dimensions of the test pipe (wall thickness 8.1 mm and 7.2 mm respectively). Concerning the validity, it was additionally explained that with a variation of the decisive conditions, the registration degrees are invalid. The specific change rate registration degrees were not identified for other parameter configurations. In this paper, specific change rate registration degrees are determined for the variation of the medium temperature change rate and the pipe wall thickness. (authors)