Abstract Scope |
In-service welding in pipelines can be applied for repair and fittings installation purposes. One of the most important things in such application is the determination of the welding parameters, avoiding issues as hydrogen induced cracks (HIC), blow-out and fragile microstructures. Several different methods of in-service welding can be found, but this paper focus on the orbital beads carried out to seal the repairs by sleeve type-B. These orbital passes can face severe cooling rates depending on the flowing inside the pipe and consequently heating exchange between the pipe’s wall the fluid, promoting susceptible conditions for welding defects and risks aforementioned. Thus, this paper presents the development of a welding procedure using the pulsed GMAW and a Finite Element Model (FEM) to predict thermal cycles, cooling rates and the final microstructure for in-laboratory experiments. The application of inductive heating is also approached and modeled as an alternative to conventional heating methods. The model was created based on real experiments using a water looping to simulate real conditions found in the field. The predicted thermal cycles are validated based on the welding thermal cycles measured by thermocouples a priori.
The development of the experimental tests was carried out in a water looping. The looping consists of a pump and a tube circuit connected to a carbon steel pipe with 2 m length. The water temperature was controlled at 20 °C using a chiller. For the inductive heating it was applied a Miller Proheat 35 kW power source. The base material used in this paper was an API 5L grade B with 12” of outside diameter. The filler metal was an ER 70S-6 with 1.2 mm in diameter.The simulated sleeve was made using the same base material and cut in two halves, wrapping the main pipe. The welding process was the GMAW with pulsed current.
The FEM simulation was done using the COMSOL Multiphysics due to its capability to couple different physics interfaces. The model runs in two steps according to the sequence developed in the real applications. This means that in the first step (preheating) the model solves the induction physic, heating up the sleeve and the pipe simultaneously, simulating a preheating condition. Then, in the second step (welding) the heat transfer is solved keeping the induction heating from the first step.
Pulsed GMAW process provided a very stable molten pool, for all positions in orbital welding. The cross section macrograph did not show welding defects and its correlation with the macrographs highlighted regions with high hardness microstructure. The most critical point in such aspect was the coarse grain heat affected zone (CGHAZ) of the second pass hitting up to 390 HV. A microstructural analysis in the CGHAZ showed the presence of martensitic and bainitic structures precipitated due to the fast cooling rate caused by the water flow inside the pipe. Images showed both microstructures formed inside the large grains. Such microstructures can severely compromise the component’s performace, because HIC can occur in such zones. Its formation is fundamentally dependent on the in-service parameters as welding heat input and flow conditions, enphasizing the importance of a model to evaluated preminarly to the final weld.
Simulated molten zone for all three beads which compared with the real molten zone (macrograph) presented an error maximum of 10%. In order to validate the thermal cycles for each pass, a probe was positioned at the same distance from the weld bead measured after each orbital pass and the maximum temperature at that point presented a maximum error of 1%. In the tail out of the thermal cycles, which represents the cooling rate, the maximum error was 10% in the temperature values.
• Orbital welding of sleeves type-B for in-service application using the GMAW with pulsed current provided excellent weld beads without defects, but in the microstructural analysis it was identified the presence of fresh martensitic and bainitic structures;
• The model developed was able to meet the molten zone and the welding thermos cycles with an accuracy of 1% in the peak temperature and 10% of error maximum in the tail out (cooling phase);
• The model developed simulate all the steps for welding in-service, including the preheating phase using the inductive heating.
Key-words: Modeling; Repair welding; Sleeve type-B; Thermal cycles; Martensite.
|