Sandvik Sanicro 25, A new material for ultrasupercritical coal fired boilers.

R.Rautio S.Bruce

Sandvik Materials Technology SE-811 81 Sandviken

Sweden

Sandvik Sanicro 25, A new material for ultrasupercritical coal fired boilers.

R.Rautio S.Bruce

Sandvik Materials Technology SE-811 81 Sandviken

Sweden

Abstract

The power generation industry worldwide has identified a need for the development of coal-fired boilers operating at much higher efficiencies than the current generation of supercritical plants. This increased efficiency is expected to be achieved using ultrasupercritical steam conditions, which will themselves require the development of new materials for critical components. In order to limit as far as possible the use of alloys using expensive alloying materials it will be necessary to maximise the utilisation of the strength and corrosion capabilities of materials throughout the spectrum of materials from ferritic through austenitic to nickel based materials. Sandvik Materials Technology has developed an austenitic alloy, Sanicro 25, with excellent high temperature strength and corrosion resistance utilising an economical compositional balance of alloys. The alloy has been developed for use within 700°C (1300°F)/ 300 bar (4500psi) steam conditions and is an excellent class leading candidate material for such high temperature applications. This paper provides an introduction to the alloy, its development status and properties.

1. Alloy design concept.

The first phase of the project included a conceptual study which comprised a review of published information and practical experience of steelmaking, primary processing and tubemaking.

After the conceptual study the objectives were defined. The requirements for the new austenitic grade for ultrasupercritical coal-fired boilers being as follows: high creep and oxidation resistance, high microstructural stability and good fabricability being necessary properties for the new grade.

The various objectives for the properties of the new material required the use of a number of basic alloy design principles.

The main technique employed to reach the very high elevated temperature strength targeted is precipitate strengthening through utilisation of Nb(C,N), fine NbCrN, M23C6 and Cu-rich precipitates. The high temperature strength is also improved by the addition of W. In order to improve the hot corrosion resistance a Cr content of approximately 22,5-24,0wt% together with a low Mo content has been used.

In order to ensure a balanced composition consideration has been given to the effects of the high Cr and W levels which promote formation of sigma phase. Thus the Ni content has been increased to approximately 25.0% together with a relatively high amount of nitrogen. For similar reasons the levels of Mo and Si have been kept relatively low.

2. Experimental manufacture

With the above design principles in mind a series of 13 experimental compositions were designed.

A small melt was produced for each of the 13 development compositions and one melt for the reference heat 14, the latter being based on grade NF709. The chemical composition of the 14 heats is shown in table 1.

A suitable forging temperature had to be predicted for all the experimental melts, this being done by hot ductility testing using a Gleeble machine.

Prior to hot ductility testing all of the specimens were heat treated for 30 minutes at 11000C (2000°F) followed by water quenching. From the test results the optimum hot forging temperatures were predicted.

All ingots were processed by forging to Ø87 mm bar followed by extrusion to Ø29 mm bars in the laboratory pilot scale press.

After extrusion a heat treatment test was performed. Six specimens from all charges were exposed to three different temperatures; 1100°C (2000°F), 1150°C (2100°F) and 1200°C (2200°F) respectively for two different times. The evaluation of the heat treatment was performed with regards to grain size and amount of precipitates.

Based on this evaluation it was decided to heat treat the experimental grades at 1150-1200°C (2100-2200°F) followed by water-cooling.

The extruded bars with a length of 2,2m were pickled and straightened. After straightening the bars were cold drawn which was performed in three passes without intermediate heat treatments between the drawing passes. The final dimension for the bars was Ø16,6mm. This resulted in approximately a 50% reduction of area.

Prior to heat treatment of the drawn bars a thermo-calc calculation was performed for the purpose of predicting relevant heat treatment temperatures. Heat treatment tests were performed on several combinations of time and temperature.

Table 1. Chemical composition [wt%] of the 14 experimental heats

Heat C Si Mn Cr Ni Mo W Co Cu Nb N

1 0,08 0,2 0,5 22,1 24,9 <0,01 3,0 0,0 2,9 0,32 0,18

2 0,08 0,2 0,5 22,6 25,1 <0,01 3,5 0,0 3,0 0,34 0,22

3 0,08 0,3 0,5 22,5 25,0 <0,01 2,2 0,0 3,0 0,42 0,21

4 0,08 0,2 0,4 22,4 25,1 <0,01 2,3 1,5 3,0 0,44 0,20

5 0,08 0,2 0,5 22,0 25,0 <0,01 3,5 0,0 3,0 0,42 0,21

6 0,08 0,2 0,5 22,2 24,9 <0,01 3,5 1,5 3,0 0,49 0,23

7 0,08 0,2 0,5 22,5 25,1 <0,01 2,3 0,0 2,9 0,32 0,20

8 0,08 0,2 0,5 22,6 25,1 <0,01 3,0 0,0 3,0 0,43 0,25

9 0,08 0,2 0,5 23,9 26,6 <0,01 2,2 0,0 3,0 0,33 0,22

10 0,08 0,2 0,5 24,0 26,5 <0,01 2,2 1,5 3,0 0,47 0,23

11 0,08 0,2 0,5 22,6 25,1 <0,01 2,3 0,0 0,0 0,35 0,21

12 0,08 0,2 0,5 22,6 25,1 <0,01 2,2 0,0 0,0 0,34 0,21

13 0,08 0,2 0,5 22,4 25,0 <0,01 3,1 0,0 0,0 0,41 0,22

14* 0,08 0,5 1,1 20,0 24,9 1,5 0,0 0,0 0,0 0,27 0,20

* Reference material being based on NF709.

2.1 Evaluation of experimental heats

The experimental heats were evaluated with a short screening tests of tensile, impact, creep strength, and fireside corrosion.

The screening tensile tests of the 14 heats was performed by Centro Sviluppo Materiali S.p.A. (CSM) at the following temperatures 20°C (70°F), 400°C (700°F), 500°C (930°F), 600°C (1100°F) and 700°C (1300°F). The data indicated that the tensile properties of the experimental alloys was similar to, or better than reference heat 14 (1).

Ageing was performed at 700°C (1300°F) for 1000h , 3000h and 10000h, followed by impact testing of aged and non-aged material. The impact values indicate that heat 4, 12, 13 and 14 have started to form stable carbides and nitrides after 1000h of ageing and after 10000h of ageing all of the heats have formed stable carbides and nitrides (1).

The 14 heats, 13 experimental plus one reference heat have been subjected to screening for creep rupture properties. The test temperatures were 6500C (1200°F) and 7000C (1300°F), with stress levels selected to aim for 1000 and 3000 hours test duration. The test parameters were chosen with regard to creep performance of the material grade NF709. Stress of 255MPa (37ksi) for the 1000h testing and 220MPa (32ksi) for the 3000h testing at 650°C (1200°F), and a stress of 165MPa (24ksi) for the 1000h testing and 140MPa (20ksi) for the 3000h testing at 700°C (1300°F) ) were selected. The tests at 6500C (1200°F) were performed by Joint Research Centre (JRC) and the tests at 7000C (1300°F) by CSM. The creep test results of the experimental heats turned out to be very good. Heat 6 having the longest creep rupture times being five times longer than for the reference material, heat 14 (1).

In order to screen the experimental heats for fireside corrosion a number of heats 2, 6, 8 and 13 have been fireside corrosion tested by Ansaldo Ricerche s.r.i at the temperatures 700°C (1300°F) for 5000h. The test specimens were exposed for coal ash and a continuous gas flow during the test. The ash contained Na2SO4, K2SO4, Fe2O3, Al2O3 and SiO2 and the gas contained SO2 and CO2. Samples were removed from the furnace at several intervals during the test and sample weight change determined. The test results indicated that all the tested materials suffered slight corrosion. Materials with the higher Cr and Ni contents seemed to show the best performance in the simulated environment (1).

3. Materials qualification

Following the screening tests heat 6 was selected from the screening tests, because of the superior creep strength and good mechanical properties, see table 1 for the chemical composition of the selected heat.

As part of the evaluation of the selected alloy, manufacture of further test material in tubular form was required. The tube size selected for testing was Ø41,4x8 mm average wall in cold pilgered and annealed finish. A new melt based on the chemical composition for heat 6 was produced, see table 2 for chemical composition for heat 15. The selected alloy was successfully produced by extrusion to Ø73,0x11,0 mm followed cold pilgering to the final dimension of Ø41,4x8 mm followed by solution annealing. See figure 1 for tube sample.

Table 2. Chemical composition [wt%] of the produced tubes of Sanicro 25

Heat C Si Mn Cr Ni Mo W Co Cu Nb N

15 0,08 0,25 0,51 22,6 25,5 0,02 3,45 1,57 2,99 0,44 0,24

Figure 1. Sanicro 25 tube Ø41,4x8 mm

3.1 Fabricabillity tests

3.1.1 Bending

Four different bends were performed on the produced tubes by Mitsui Babcock Energy Limited (MBEL). Two different 180° bends and two different 90° bends were performed, the bending radius ratio were 3,1 and 2,3 for both of the 180° and 90° bends. All the bends were performed at 20°C (70°F). The bending operation was performed successfully without any cracks in the tube wall and the maximal measured ovality on the tube was 1.6% (1), see figure 2 and figure 3 for a sample of the bends.

Figure 2. Two different 180° bends.

Figure 3. Two different 90° bends.

3.1.2 Welding

An alternative to weld Sanicro 25 is to use the filler metal alloy 617 which is currently the first choice of available products to use as filler metal for Sanicro 25 welds. The literature data for alloy 617 indicates that alloy 617 has satisfying corrosion and mechanical data (2). Creep testing of welded tubes with filler metal alloy 617 is in progress. The creep test is being performed at 650°C (1200°F) and 700°C (1300°F) with stress levels selected to give estimated rupture times of 1000h, 3000h and 10000h. Early test data indicate that 617 could be a suitable filler metal but more test time is required in order to verify suitability.

The long-term solution and most sustainable solution for filler material is probably to develop a matching Sanicro 25 filler. This solution would include manufacturing of filler material with the same composition as the parent material Sanicro 25, as well as welding and creep trials.

3.2 Creep rupture

Creep rupture tests has been carried out by CSM, JRC and Electricite de France (EDF) at 650°C (1200°F), 700°C (1300°F) and 750°C (1390°F) on both bar and tube material.

The main target for this new austenitic grade for ultra supercritical boilers is a creep strength of 100MPa (14,5ksi), 700°C (1300°F) at 100000 hours. With the main target in mind the stresses for the creep tests at 650°C (1200°F) and 700°C (1300°F) were designed to give estimated test duration time to fracture of 1000, 3000, 10000, 30000 and 70000 hours. Stresses for test at 750°C (1390°F) were designed to give shorter times than for the other two temperatures.

Ongoing creep rupture tests with test times above 25000 hours indicated that the mechanical target 100 MPa (14,5ksi), 700°C (1300°F) at 100000 hours will be achieved for the bar material, see figure 4 for master curve, the blue points indicates broken creep rupture tests for bar material from heat 6.

Test data also indicate that the mechanical target 100 Mpa (14,5ksi), 700°C (1300°F) at 100000 hours could be achieved for tube material. However, more test time is needed to state that the target will be achieved. CSM has done, using the data from test times up to 5000 h, estimation for the creep rupture time on the tube material to be between 70000-100000 hours for 100MPa (14,5ksi) at 700°C (1300°F). It can be noticed that Sanicro 25 has notable longer creep rupture time compared to the litterature data for HR3C (3) and NF709 (4) at 700°C (1300°F), see figure 5.

The master curve for Sanicro 25 is shown in figure 3. The red and the green spots in the master curve indicate broken creep rupture tests for tube material from heat 15. The yellow spots indicate running creep tests in January 2004.

Figure 4. Sanicro 25 master curve.

100

110

120

130

140

150

160

170

180

190

200

100 1000 10000 100000TIME [hours]

STR

ESS

[MPa

]

HR3C (Sumitomo) NF709 (Nippon Steel)Sanicro 25 (ABSMT) Sanicro 25 (Running)

NF709HR3C Sanicro 25

Creep Rupture Strength 700°C

Figure 5. Creep rupture strength 700°C.

3.3 Tensile properties

Tensile tests on Sanicro 25 from heat 15 have been performed at the temperature range from 20°C (70°F) to 800°C (1500°F) with two specimens tested at each temperature. See figure 6 for the tensile properties.

0

100

200

300

400

500

600

700

800

900

20 50 100 150 200 250 300 350 400 450 500 550 600 650 700 750 800Temperatur °C

Stress,MPa

0.2% PS UTS

Figure 6. Tensile properties for Sanicro 25

3.4 Fatigue test

Powergen have tested Sanicro 25 and two commercially materials, SAVE25 and NF709 for low cyclic fatigue.

All three materials were tested in the as-received condition for low cyclic fatigue test, additionally Sanicro 25 and SAVE25 were also tested in the aged condition with Sanicro 25 being aged at 700°C (1300°F) for 2000h and SAVE25 aged at 700°C (1300°F) for 3000h. All the materials gave essentially similar results in the LCF test, see figure 7.

Tensile dwells had a significant effect on reducing endurance in both SAVE25 and Sanicro 25 compared with continuous cycling. However, when examined in terms of the plastic strain range during testing the dwell and continuous cycling data fall on the trend line for each material, figure 8. This would suggest that a five minute dwell has little effect in terms of creep damage contribution but merely modifies the stress-strain

loop by increasing the plastic strain contribution via stress relaxation in the dwell (5), figure 8.

0,3

0,35

0,4

0,45

0,5

0,55

0,6

0,65

0,7

0,75

1000 10000 100000Number of Cycles to crack Initiation

Tota

l Str

ain

Ran

ge (%

)

NF709 SAVE25 Save 25 5 min dwellSanicro 25Save 25 AgedSave 25 Aged 5 min dwellSanicro 25 5 min dwellSanicro 25 aged

Figure 7. LCF data – total strain range

0,01

0,1

1

1000 10000 100000

Number of Cycles to Crack Initiation

Plas

tic S

tran

Ran

ge (%

)

NF709SAVE25SAVE25 DwellSanicro 25Sanicro 25 dwellSAVE25 AgedSAVE25 Aged DwellSanicro 25 Aged

Figure 8. LCF data – plastic strain range

3.5 Microstructural examination

A microscopic examination was performed of tubes from heat 15 and of bar from heat 6. Samples in the as delivered condition, aged for 1000h at 700°C (1300°F) and aged for 3000h at 700°C (1300°F) were examined using the scanning electron microscope (SEM), figure 9. The samples aged for 1000h at 700°C (1300°F) were also examined using transmission electron microscope (TEM) (6), figure 10.

These examinations showed that samples in the as delivered condition had Nb-rich precipitates which were in addition enriched with Cr. The aged samples contained small needle-shaped M23C6 in both the grains and grain boundaries. Sigma-phase and Cr2N were not detected in the samples aged for 1000h and 3000h. The grain size for the tubes was 5,0 ASTM and the grain size for the experimental bar was 5,5 ASTM (7).

Figure 9. SEM image, microstructure with Nb-rich preicipitates.

Figure 10. TEM image, spherical and needle-like M23C6 particles.

3.6 Hot corrosion

Hot corrosion tests were carried out at 700°C (1300°F) for an exposure time of 3000h by Ansaldo ricerche s.r.i. NF709 was chosen as a reference material because of NF709 is known as one of the best commercially available steels in this respect.

Test specimens of both NF709 and Sanicro 25 was covered with ash and exposed for a gas flow during the test time. The ash composition was 5% Na2SO4, 5% K2SO4, 30%

Fe2O3, 30% Al2O3 and 30% SiO2 and gas composition was 0.25% SO2, 3.5% O2, 15% CO2 and balance of N2. The samples were covered by a definite amount of ashes in a ethil alchohol suspension, ashes were renewed every 100h for the first 500h of testing and then every 500h up to the end of the test (3000h). This methodology assures a better exchange between gas environment and metal surface and consequently a better simulation of the material performance in service.

After exposition corroded samples were subjected to a chemical descaling process in order to determine the mass change for each material as a function of test duration. Figure 11 shows the results obtained for the two materials studied. It is clear from the graph that Sanicro 25 has much lower corrosion rate than NF709. Also corrosion morphology differs a lot from one material to the other one, Sanicro 25 is characterised by localised corrosion with some deep pits especially after longer time of exposition, NF709 shows a more uniform corrosion front with a great loss of thickness after 3000h of exposition (8). Figure 12 and 13 shows microstructure of the two different grades after 2000h of exposition (8).

Figure 11. Weight change after descaling as function of test duration.

Figure 12. Microstructure of Sanicro 25 after 2000h of exposition.

Figure 13. Microstructure of NF709 after 2000h of exposition.

4. Summary

A conceptual study resulted in 14 trial casts of material with varying compositions considered as having potential to meet the targets of 100 000 hour creep rupture strength of 100MPa (14,5ksi) at 7000C (1300°F) together with possessing fireside corrosion resistance better than NF709.

Superheater tube of dimension Ø41,4 x 8 mm was successfully produced, with welding and bending operations on the tube executed showing good results demonstrating the material has good fabricability.

Ongoing mechanical testing on trial cast material in bar form with testing times above 25000 hours indicates that the mechanical target 100MPa (14,5ksi), 700°C (1300°F), 100000 hours will be achieved.

Test data from ongoing mechanical testing of tube material indicate that the mechanical target 100 MPa (14,5ksi), 700°C (1300°F) at 100000 hours could be achieved. However, more test time is needed to state that the target will be achieved for tube material. Using the data available so far it can be stated that the creep rupture time of the tube material will be between 70000-100000 hours for 100MPa (14,5ksi) at 700°C (1300°F). The tests also shows that Sanicro 25 has notable longer creep rupture time compared to NF709 at 700°C (1300°F).

References

1. Internal Thermie project database, 2003. (The database is open only for participants in Thermie project).

2. Special metals datasheet for Alloy 617, 2004-06-29, http://www.specialmetals.com/products/inconelalloy617.htm

3. Development of tubes and pipes for ultra-supercritical thermal power plant boilers. H. Naoi, H. Mimura: Nippon steel technical report, No. 57 April 1993.

4. K. Yoshikawa, Y. Sawaragi, and H. Yuzawa, “Development of new boiler tubes with high elevated temperature strength and high corrosion resistance”, Presented at the first international conference on improved coal-fired power plants, Palo alto, CA (November 1986).

5. Austenitic superheater an reheater materials cyclic mechanical properties, P. Mulvihill: Powergen, 2002.

6. Microscopic examination of laboratory manufactured austenitic heats for the Thermie project, A. sundström: Sandvik Materials Technology, June 2000. T000843.

7. Microscopic examination of austenitic tubes for the Thermie project, R. Rautio: Sandvik Materials Technology, October 2003. 031123TE.

8. Hot corrosion tests on austenitic steels, Ansaldo Ricerche s.r.i, June 2002. R04610UV6000L.