posted on 2024-11-12, 09:10authored byElizabeth Budzakoska-Testone
Repair rather than replacement of pressure vessel components operating at elevated temperatures is becoming a far more viable option, with welding being used increasingly for repair, replacement, retrofits and modifications. However, the considerable cost and time involved in performing conventional post weld heattreatment (PWHT) is increasingly forcing utility owners to turn toward other alternatives, such as cold weld repairs. These require no PWHT and rely on a controlled deposition process – precise weld bead placement and heat inputs to achieve tempering of the HAZ. One such cold weld repair was applied to a prematurely failed reheater outlet header (RHOH) in an emergency situation, enabling its return to service until a replacement component was available. Investigation of this particular repair constitutes a core component of this thesis. Much of the research that has been conducted to justify cold repair techniques has used accelerated high temperature creep testing to demonstrate their integrity. How well this testing reflects real-life performance of repaired components remains uncertain. However, the present study provided a rare opportunity to evaluate the effects of service exposure on the performance of an emergency cold weld repair, as well as to examine the effectiveness of accelerated laboratory creep testing on creep life predictions for the repair weld, the original fabrication welds and the service exposed parent metal. The metallurgical and mechanical properties of the service-exposed non-PWHT temper bead (TB) repair weld were analysed and characterised systematically relative to both the service-exposed original fabrication weld and parent metal. All test specimens were machined transverse to the welding direction, from 20 mm thick slices taken from both the TB repair and original fabrication weldments. The mechanical property characterisation involved hot tensile (miniature sample), Charpy V-notch impact, hardness and accelerated creep-rupture testing. The microstructural characterisation consisted of compositional analysis, surface replication and conventional metallography, post-test fractography, high-resolution microscopy (SEM, TEM and EBSD), EDS and dilatometry. The microstructural studies were successful in characterising the service-exposed TB repair weld metal both in the as-received and post-test conditions. The findings generally showed the TB repair weld metal microstructure was significantly different from those of the parent and fabrication weld metals. This was established by the TEM and EBSD studies which showed finer carbides, higher dislocation density and limited subgrain formation, features which are consistent with structure of high stored strain energy. The finer carbides, significant Fe content of some carbides and the presence of Mo-rich carbides are also consistent with a relatively early stage in the carbide evolution sequence for this steel. Additionally, the non-metallic inclusions, which are generally considered to be deleterious to properties, were fine scale and uniform in size distribution. These microstructural characteristics were reflected in superior mechanical properties for all tests, with the notable exception of creep performance. In terms of creep resistance, weld repairs of this kind install a high energy microstructure (finer carbides, higher dislocation density, fine grain structure and subgrain / cell structure) into significantly creep degraded material and generate sharp gradients in microstructure and properties. Overall, the microstructural observations in the service-aged condition and the mechanical properties determined by the experimental work were found to support the conclusion that the ex-service material had clearly not exhausted its creep service-life. This was proven by simulating further service life using accelerated creep testing. During these experiments, the creep failure mode was found to be ductile fracture induced by micro-void coalescence. The tests indicated that although the predicted creep life of the TB repair weldment was inferior to the aged parent metal, it was slightly higher than that of the service-exposed fabrication weldment. It is therefore concluded that the TB repair weld and its associated HAZ did not compromise the overall weldment integrity. This conclusion is drawn despite the finding that the accelerated creep test data for the repair weld showed sub-normal creep performance relative to NT samples and even the aged parent metal. This outcome is probably due to the “accelerated” test conditions which induced structural changes that are unlikely to occur under the actual service conditions. Therefore, accelerated creep test data should be treated with circumspection, particularly for structures such as cold repair weld metal which are characterised by high stored strain energy. The instability of this kind of structure in relation to restoration processes (at the “accelerated” temperatures used for testing) is likely to produce premature creep failure and result in a highly conservative creep life assessment. The perceived danger of cold repair welding is that the vessel integrity will be impaired by the absence of a PWHT to reduce internal stresses, liberate hydrogen and improve mechanical properties. However, the evidence obtained in this study indicates that this need not be the case. Therefore, it is concluded that the TB weld technique can be successfully applied to high energy piping (HEP) components without the need for PWHT. The results of the investigation provide new insights into the microstructure and properties, particularly the creep performance, of service-aged cold repair welds. The research has also revealed a serious limitation of accelerated creep testing of weld repair structures of this kind, in that dynamic restoration processes can be induced that are unlikely to occur at the lower service operating temperatures and stresses.
History
Year
2014
Thesis type
Doctoral thesis
Faculty/School
School of Mechanical, Materials and Mechatronic Engineering
Language
English
Disclaimer
Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.