CREEP BEHAVIOUR AND MICROSTRUCTURE OF A HETEROGENEOUS P23...

7
METAL 2009 19. – 21. 5. 2009, Hradec nad Moravicí ___________________________________________________________________________ 1 CREEP BEHAVIOUR AND MICROSTRUCTURE OF A HETEROGENEOUS P23/P91 WELD Vlastimil Vodárek Lucie Střílková Zdeněk Kuboň MATERIÁLOVÝ A METALURGICKÝ VÝZKUM s.r.o., Pohraniční 31, 706 02 Ostrava- Vítkovice, ČR, [email protected] Abstract The structural integrity of welded components operated at elevated temperatures is of key importance in power plant applications. Long-term creep exposure of heterogeneous welds is accompanied by redistribution of interstitial elements which strongly affects microstructural evolution in the vicinity of the fusion zone between the low and high alloy materials. The paper summarises results of studies on creep rupture properties and minor phase evolution in the P23/P91 heterogeneous weld during creep exposure at 500, 550 and 600°C. Results of creep rupture tests on the cross weld specimens were close or slightly below the lower limit of the ±20% scatter band around the standardized curve for creep strength of the P23 steel. In the course of creep exposure the development of creep damage could simultaneously take place in several critical areas of the weldment but the location of the final failure was determined by the “weakest area” for given testing parameters. Experimental data on microstructural evolution have been matched with results of thermodynamic and kinetic simulations. Results of studies on precipitation reactions in the individual weld zones have mostly been in accordance with the Thermocalc and Dictra simulations, except the P23 base material. It has been demonstrated that undissolved MX particles in the partly decarburized zone of the P23 steel significantly delay recrystallization of the bainitic matrix. 1. Introduction Creep strength of heterogeneous welds is strongly dependent on the microstructural evolution in individual weld zones during creep exposure [1]. A very important mechanism of microstructural degradation of dissimilar welds represents “up-hill” diffusion of interstitial elements [2]. Decarburization during exposure at elevated temperatures can cause progressive weakening of a low alloy steel adjacent to the fusion boundary with a high alloy steel. On the other hand the carburized layer on the opposite side of the fusion boundary is much stronger than the surrounding matrix. In nitrogen-bearing steels redistribution of nitrogen is important too. Uniaxial cross weld specimens of dissimilar welds often fail by artificial mechanisms involving complex strain interactions between adjacent weld zones [3]. Adjacent strong material can constrain deformation and retard failure in a weak zone which is thin compared to the specimen diameter. Prediction of real weld behaviour requires detailed evaluation of differences between laboratory tests and power plant components. Progress in thermodynamic and kinetic modelling makes it now possible to simulate microstructural evolution in heat resistant steels and their weld joints during long-term exposure. The reliability of modelling depends strongly on the quality of databases. That is why experimental investigations are required to verify results of microstructural modelling.

Transcript of CREEP BEHAVIOUR AND MICROSTRUCTURE OF A HETEROGENEOUS P23...

Page 1: CREEP BEHAVIOUR AND MICROSTRUCTURE OF A HETEROGENEOUS P23 …konference.tanger.cz/data/metal2009/sbornik/Lists/Papers/051.pdf · CREEP BEHAVIOUR AND MICROSTRUCTURE OF A HETEROGENEOUS

METAL 2009 19. – 21. 5. 2009, Hradec nad Moravicí ___________________________________________________________________________

1

CREEP BEHAVIOUR AND MICROSTRUCTURE OF A

HETEROGENEOUS P23/P91 WELD

Vlastimil Vodárek Lucie Střílková Zdeněk Kuboň

MATERIÁLOVÝ A METALURGICKÝ VÝZKUM s.r.o., Pohraniční 31, 706 02 Ostrava-

Vítkovice, ČR, [email protected] Abstract The structural integrity of welded components operated at elevated temperatures is of key importance in power plant applications. Long-term creep exposure of heterogeneous welds is accompanied by redistribution of interstitial elements which strongly affects microstructural evolution in the vicinity of the fusion zone between the low and high alloy materials. The paper summarises results of studies on creep rupture properties and minor phase evolution in the P23/P91 heterogeneous weld during creep exposure at 500, 550 and 600°C. Results of creep rupture tests on the cross weld specimens were close or slightly below the lower limit of the ±20% scatter band around the standardized curve for creep strength of the P23 steel. In the course of creep exposure the development of creep damage could simultaneously take place in several critical areas of the weldment but the location of the final failure was determined by the “weakest area” for given testing parameters. Experimental data on microstructural evolution have been matched with results of thermodynamic and kinetic simulations. Results of studies on precipitation reactions in the individual weld zones have mostly been in accordance with the Thermocalc and Dictra simulations, except the P23 base material. It has been demonstrated that undissolved MX particles in the partly decarburized zone of the P23 steel significantly delay recrystallization of the bainitic matrix. 1. Introduction Creep strength of heterogeneous welds is strongly dependent on the microstructural evolution in individual weld zones during creep exposure [1]. A very important mechanism of microstructural degradation of dissimilar welds represents “up-hill” diffusion of interstitial elements [2]. Decarburization during exposure at elevated temperatures can cause progressive weakening of a low alloy steel adjacent to the fusion boundary with a high alloy steel. On the other hand the carburized layer on the opposite side of the fusion boundary is much stronger than the surrounding matrix. In nitrogen-bearing steels redistribution of nitrogen is important too. Uniaxial cross weld specimens of dissimilar welds often fail by artificial mechanisms involving complex strain interactions between adjacent weld zones [3]. Adjacent strong material can constrain deformation and retard failure in a weak zone which is thin compared to the specimen diameter. Prediction of real weld behaviour requires detailed evaluation of differences between laboratory tests and power plant components. Progress in thermodynamic and kinetic modelling makes it now possible to simulate microstructural evolution in heat resistant steels and their weld joints during long-term exposure. The reliability of modelling depends strongly on the quality of databases. That is why experimental investigations are required to verify results of microstructural modelling.

Page 2: CREEP BEHAVIOUR AND MICROSTRUCTURE OF A HETEROGENEOUS P23 …konference.tanger.cz/data/metal2009/sbornik/Lists/Papers/051.pdf · CREEP BEHAVIOUR AND MICROSTRUCTURE OF A HETEROGENEOUS

METAL 2009 19. – 21. 5. 2009, Hradec nad Moravicí ___________________________________________________________________________

2

2. Experimental Materials and Techniques The weld joint made of P23 and P91 steel pipes of the dimensions Ø219x25mm were fabricated in SES Tlmače, Slovakia [4]. A P91 matching filler metal (E CrMo 9 1B, ø 3.2 and 4.0mm) was applied. The chemical compositions of individual materials in the weldment investigated are stated in Table 1.

Table 1. Chemical compositions of base materials and the weld metal (WM), wt.% Steel C Mn Si Ni Cr Mo V Ti Nb W N

P23 0.08 0.55 0.27 0.08 2.11 0.07 0.23 0.06 0.01 1.70 0.013

P91 0.11 0.51 0.38 0.42 8.67 1.00 0.23 0.01 0.07 0.01 0.048

WM91 0.11 0.66 0.21 0.82 9.50 1.02 0.22 0.01 0.04 0.06 0.028

Cross weld creep specimens included base materials, heat affected zones and the weld metal. Uniaxial creep rupture tests have been performed at temperatures 500, 550 and 600°C in the stress interval from 55 to 200 MPa. Failure locations were identified using optical metallography. Furthermore, hardness profiles were evaluated on longitudinal sections through the ruptured cross weld specimens. EBSD investigations (COM) in the WM91/P23 fusion zone were carried out on a Sirion 200 FEG SEM equipped with HKL Technology Channel 5 System. Modelling of the thermodynamic equilibrium was carried out using the ThermoCalc software and the STEEL 16 database [5]. Calculations of carbon redistribution and phase profiles across the P23/P91 interface for 500, 550 and 600°C/30 000 hours were performed using the Dictra software [5]. Investigations on minor phases in creep ruptured specimens after approx. 30 000hours were carried out on a JEOL JEM 2100 equipped with a PGT EDX analyser. Minor phases were identified by both EDX and SAED techniques. Carbon extraction replicas were prepared in the following regions of the creep ruptured specimens:

� base materials P23 and P91, � decarburized zone of the P23 steel, � carburized zone of the WM91.

3. Results and Discussion Data of creep rupture tests are summarised in Fig. 1. Still ongoing tests are marked by empty symbols. Together with the experimental results, the mean and minimum (-20% of the mean value) standardized creep rupture strength curves for the P23 steel are shown. So far available results of creep rupture tests are close or slightly below the –20% creep strength curve. The most pronounced drop of creep resistance occurred at the lowest applied stresses and 600°C. Metallographic investigations proved that preferred locality of creep failure was a complex function of testing parameters and local microstructural characteristics, Table 2. Three critical localities prone to the development of creep damage were identified [6]:

� intercritical area (IC) of HAZ in the P23 steel, � fusion zone on the side of the P23 steel. In this coarse grain area (CG) of HAZ partial

decarburization took place during the both PWHT and creep exposure. � intercritical and/or fine grain (FG) regions of HAZ in the P91 steel.

In the course of creep exposure the development of creep damage could simultaneously take place in several critical areas of the weld but the location of the final failure was determined by the “weakest area” for given testing parameters. Fig. 2 shows the location of failure in the specimen ruptured after creep exposure 600°C/5 171 hours. It failed in the coarse grain HAZ

Page 3: CREEP BEHAVIOUR AND MICROSTRUCTURE OF A HETEROGENEOUS P23 …konference.tanger.cz/data/metal2009/sbornik/Lists/Papers/051.pdf · CREEP BEHAVIOUR AND MICROSTRUCTURE OF A HETEROGENEOUS

METAL 2009 19. – 21. 5. 2009, Hradec nad Moravicí ___________________________________________________________________________

3

of the P23 steel. This location is expected to be weak due to redistribution of interstitial elements during both PWHT and creep exposure. However, cavitation damage in this specimen was also revealed in the intercritical HAZ of the P91 steel, Fig. 3.

10

100

1000

21000 22000 23000 24000 25000 26000 27000

PLM=T(25+log t)

Str

es

s (

MP

a)

500°C

550°C

600°C

mean P23

low P23

Fig. 1. Results of creep rupture tests Table 2. Creep test parameters and failure locations

Temperature

[°C]

Stress

[MPa]

Time to

rupture [h]

R.A.

[%]

Failure

locations

500 200 18 802 50.8 BM P23

500 190 31 332 15.9 IC HAZ of P23

550 150 9 715 26.3 IC HAZ of P23

550 140 8 665 4.1 FL WM/P23, CG HAZ of P23

550 125 23 961 4.1 FL WM/P23, CG HAZ of P23

550 115 26 386 8.2 CG HAZ of P23

600 110 1 384 2.1 CG HAZ of P23

600 100 4 056 1.8 FG HAZ of P91

600 90 5 171 8.2 FL WM/P23, CG HAZ of P23

600 75 9 852 0.6 FG HAZ of P91

600 55 26 780 2.4 CG HAZ of P23

(and ~ 85% FG HAZ of P91)

where FL WM/P23 is a fusion line between WM and P23 base metal.

Page 4: CREEP BEHAVIOUR AND MICROSTRUCTURE OF A HETEROGENEOUS P23 …konference.tanger.cz/data/metal2009/sbornik/Lists/Papers/051.pdf · CREEP BEHAVIOUR AND MICROSTRUCTURE OF A HETEROGENEOUS

METAL 2009 19. – 21. 5. 2009, Hradec nad Moravicí ___________________________________________________________________________

4

Fig. 2 Failure location in the specimen after creep exposure 600°C/5 171 hours

Fig. 3. Cavities in IC HAZ of the P91 steel Fig. 4. Temperature dependencies of after creep exposure 600°C/5 171 hours carbon activity for the P23 and P91 steels 3.1 Microstructural Modelling Microstructural stability of dissimilar welds is strongly affected by redistribution of interstitial elements in the course of the long-term thermal/creep exposure. Diffusion of interstitial elements is driven by the differences in activities of these elements across the fusion boundary. Atoms move in the direction of activity gradients [7]. Simulation of microstructural evolution included following issues:

� stable phases in the P23 and P91 steels as a function of temperature and carbon content,

� temperature dependence of carbon and nitrogen activity in the P23 and P91 steels, � carbon redistribution close to the P23/WM91 boundary, � phase profiles across the P23/WM91 boundary for temperatures 500, 550 and 600°C

and time of exposure of 30 000 hours. Due to significant differences in carbon activity for the P23 and P91 steels in the temperature range from 500 to 800°C pronounced redistribution of carbon from the P23 to the P91 steel is to be expected, Fig. 4. On the other hand, differences in nitrogen activity for both steels are negligible and no significant redistribution of nitrogen is awaited [8]. Calculations of carbon migration and phase profiles across the P23/P91 interface for 550°C/ 30 000 hours are shown in Figs. 5 and 6, respectively. Results of calculations demonstrate “up-hill” diffusion of carbon from the P23 to the P91 steel. Both the content of carbon and the width of the carburized zone increase with the temperature of exposure. The calculated

P91 P23 WM

Page 5: CREEP BEHAVIOUR AND MICROSTRUCTURE OF A HETEROGENEOUS P23 …konference.tanger.cz/data/metal2009/sbornik/Lists/Papers/051.pdf · CREEP BEHAVIOUR AND MICROSTRUCTURE OF A HETEROGENEOUS

METAL 2009 19. – 21. 5. 2009, Hradec nad Moravicí ___________________________________________________________________________

5

carburization of the P91 steel for 500°C is 0.18 wt%C, while for 600°C it reaches 0.39 wt.%C. In the carburized zone the molar fraction of M23C6 phase increases with the temperature of exposure, but the molar fraction of MX remains about the same and the molar fraction of Laves phase decreases. Partial decarburization of the P23 steel takes place mainly at the expense of dissolving M7C3 particles. No significant effect of decarburization on molar fractions of MX and M6X is predicted.

Fig. 5. Carbon redistribution across the Fig. 6. Minor phase profiles across the P23/P91 interface for 550°C/30 000 hours P23/P91 interface for 550°C/30 000 hours 3.2 Results of Microstructural Investigations The minor phases found in the P91 and WM91 steels are in accordance with the results of ThermoCalc calculations. In the carburized zone of the WM91 a high density of M23C6 particles was present, Fig. 7. M23C6 particles in the carburized zone were richer in iron compared to those in both the WM91 and P91 steels. Fig. 7. Precipitation of M23C6 particles in Fig. 8. Fine MX particles in the partly the carburized zone of the WM91, decarburized zone of the P23, 600°C/26 780 hours 550°C/26 386 hours

Page 6: CREEP BEHAVIOUR AND MICROSTRUCTURE OF A HETEROGENEOUS P23 …konference.tanger.cz/data/metal2009/sbornik/Lists/Papers/051.pdf · CREEP BEHAVIOUR AND MICROSTRUCTURE OF A HETEROGENEOUS

METAL 2009 19. – 21. 5. 2009, Hradec nad Moravicí ___________________________________________________________________________

6

The size of M23C6 particles in the carburized zone was very variable. However no difference in chemical composition between coarse and fine M23C6 particles was found [8]. Some differences between calculations and experimental studies were found for minor phases in the P23 steel. During quality heat treatment precipitation of MX and M7C3 phases is accompanied by precipitation of the M23C6 phase which is regarded as a metastable phase in the steel under consideration. M6C starts to form only during long-term creep exposure. The kinetics of M6X precipitation depends on both temperature and time of exposure. No M6X particles were found after exposure 500°C/31 323 hours, but shorter times of exposure at higher temperatures resulted in precipitation of M6C particles. After exposure 600°C/26 780 hours many big M6X particles were present in the matrix. These particles were predominantly situated along grain boundaries. Chemical composition of fine MX particles was very variable [8]. Coarser MX particles were rich in titanium and niobium, small MX particles were rich in vanadium. A significant amount of tungsten was dissolved in MX particles [8]. Many small MX particles were preserved in the partly decarburized zone of the specimen after 26 386 hours at 550°C, Fig. 8. In the specimen after 26 780 hours at 600°C a mixture of small MX and bigger M6X particles was present in the decarburized zone. The fraction of M6C particles in this zone was much lower compared with that in the P23 steel unaffected by carbon redistribution. On the right side of Fig. 9 a typical distribution of M6C particles (white dots, black dots are cavities) in the partly decarburized zone of the P23 steel can be seen. M6C particles are present up to the P23/WM91 interface. Neither M6X nor Laves phase particles were found in the carburized zone of the WM91 (the left side of Fig. 9). Fine MX particles preserved in the HAZ of the P23 steel slowed down recovery and recrystallization processes in the bainitic matrix. Microstructure of the partly decarburized zone in the P23 steel remained bainitic in all specimens, Fig. 10.

Fig. 9. M6C distribution in the partly Fig. 10. COM image in the fusion zone of the decarburized zone of the P23 steel (on the specimen after creep 600°C/ 26 780 hours right hand side), 600°C/ 26 780 hours 4. Conclusions Results of creep rupture tests on the P23/P91 dissimilar weld were close or slightly below the lower limit of the ±20% scatter band around the standardized curve for creep strength of the P23 steel. In the course of creep exposure the development of creep damage can simultaneously take place in several critical areas of weldments but the location of the final failure is determined by the “weakest area” for given testing parameters. It was proved that evolution of minor phases in the P23/P91 welds during creep was in a good agreement with results of microstructural modelling, except the P23 base metal. Microstructure of the partly

WM91

P23

WM91

P23

10µm

Page 7: CREEP BEHAVIOUR AND MICROSTRUCTURE OF A HETEROGENEOUS P23 …konference.tanger.cz/data/metal2009/sbornik/Lists/Papers/051.pdf · CREEP BEHAVIOUR AND MICROSTRUCTURE OF A HETEROGENEOUS

METAL 2009 19. – 21. 5. 2009, Hradec nad Moravicí ___________________________________________________________________________

7

decarburized zone in the P23 steel remained bainitic in all specimens investigated. This is related to undissolved MX (VC) particles in the partly decarburized zone which slowed down recovery / recrystallization processes in the bainitic matrix. Acknowledgement

The authors would like to express thanks to the Ministry of Education, Youth and Sport of the Czech Republic for the financial support in the frame of the projects OC099 and MSM 2587080701. The Royal Society, UK are thanked for the provision of an International Short Visits Grant for VV. Dr. S.V. Hainsworth is acknowledged for hosting the visit of VV at University of Leicester.

References [1] BHADESHIA, H.K.D.H. Design of Heat Resistant Alloys for the Energy Industries, in Proc. of Parsons 2000–Advanced Materials for 21st Century Turbines and Power Plants, IOM, London, UK, 2000, p. 1-20. [2] PILOUS, V., STRÁNSKÝ, K. Strukturní stálost svarových spojů pro energetiku, Studie ČSAV, Academia Praha, 1989 (in Czech). [3] ALLEN, D.J. Creep Performance of Dissimilar P91 to Low Alloy Steel Weldments, In Parsons 2003, IOM, London 2003, p. 281-294. [4] PECHA, J., BOŠANSKÝ, J. Welding of Similar and Dissimilar Steels, COST 522 Progress Report, SES Tlmače-WRI Bratislava, 2002. [5] ZLÁMAL, B., FORET, R., SOPOUŠEK, J. Modelování fázového složení ocelí P23 a P91 včetně jejich svarových spojů, Brno 2007 (in Czech). [6] VODÁREK, V., KUBOŇ, Z. The Creep and Fracture Behaviour of Dissimilar P23/P91 Welds, Proc. of 5th International Conference on Mechanics and Materials in Design, Porto, 2006, 267. [7] VODÁREK, V. Fyzikální metalurgie modifikovaných (9-12)%Cr ocelí, Ostrava, VŠB-TU Ostrava, 2003, p. 169 (in Czech). [8] VODÁREK, V., KUBOŇ, Z., FORET, R., HAINSWORTH, S. V. Microstructural Evolution in P91/P23 Heterogeneous Welds during Creep at 500-600°C, Proc. of the IIW International Conference Safety and Reliability of Welded Components in Energy and Processing Industry, P. Mayr, G. Posch and H. Cerjak , Eds., TU Graz, 2008, p. 233.