Academia.eduAcademia.edu

Figure 13 – uploaded by Andrea Peri

See full PDFdownloadDownload figure
Figure 13. Schematic diagram showing the oxygen diffusion process in the condensate layer. Mathematically, this is a straightforward problem, assuming the steam itself can be assumed to be oxygen-free upon start-up. The details are described in Appendix 1, which should be viewed in the context of Figure 13.

Figure 13 Schematic diagram showing the oxygen diffusion process in the condensate layer. Mathematically, this is a straightforward problem, assuming the steam itself can be assumed to be oxygen-free upon start-up. The details are described in Appendix 1, which should be viewed in the context of Figure 13.

Related Figures (58)

Figure 1. Simplified steam cycle in the power plant’.
Figure 1. Simplified steam cycle in the power plant’.
LD: Local Decision. 10,18 ‘able 1 Chemical control targets for drum-type boilers (Expressed as pg/kg)
LD: Local Decision. 10,18 ‘able 1 Chemical control targets for drum-type boilers (Expressed as pg/kg)
Stress corrosion cracking of turbine steels occurs in many locations’, as shown in Figure 6. It has been suggested'*'°”° that disc cracking occurred at or downstream of the Wilson line (locations where the steam moisture is 3 - 4%) in the turbine, suggesting that moisture (or liquid phase) was essential for cracking, whist others state*® that the early condensates (which can form at moisture of less 1%) are more important.
Stress corrosion cracking of turbine steels occurs in many locations’, as shown in Figure 6. It has been suggested'*'°”° that disc cracking occurred at or downstream of the Wilson line (locations where the steam moisture is 3 - 4%) in the turbine, suggesting that moisture (or liquid phase) was essential for cracking, whist others state*® that the early condensates (which can form at moisture of less 1%) are more important.
Figure 7 shows an example calculation for the metal temperatures at the keywa crowns of a Marchwood LP rotor and the incidence of cracking. It is clear the cracking, is closely related to the occurrence of condensation on the discs”’.
Figure 7 shows an example calculation for the metal temperatures at the keywa crowns of a Marchwood LP rotor and the incidence of cracking. It is clear the cracking, is closely related to the occurrence of condensation on the discs”’.
The distribution of temperature, pressure and wetness of the steam in the Magnox power plan pressure of Magnox sta Magnox sta s is not available. There is a large variation in the temperature and the inlet and outlet steam from station to station’’, as indicated by the ypical operating data for four stations shown in Table 7. It can be seen that the ions operate under much lower pressures compared to AGR and PWR. It has been suggested” that the temperature at which first condensation occurs at the ions is in the range 70 °C to 120 °C for the LP turbines and about 170 °C for the HP turbines. The PWR is operated under vary different conditions”, as shown in Table 6. The HP turbine is almost all wet and the Wilson line temperature is above 250 °C. The LP turbine is partly wet and the Wilson line is at less than 78 °C. It should be noted that SCC of HP turbine discs and blades has not been reported, despite condensation occurring at very high temperature. This is probably due to the fact that the loading level of HP turbine discs and blades is much lower than that of the LP turbines.
The distribution of temperature, pressure and wetness of the steam in the Magnox power plan pressure of Magnox sta Magnox sta s is not available. There is a large variation in the temperature and the inlet and outlet steam from station to station’’, as indicated by the ypical operating data for four stations shown in Table 7. It can be seen that the ions operate under much lower pressures compared to AGR and PWR. It has been suggested” that the temperature at which first condensation occurs at the ions is in the range 70 °C to 120 °C for the LP turbines and about 170 °C for the HP turbines. The PWR is operated under vary different conditions”, as shown in Table 6. The HP turbine is almost all wet and the Wilson line temperature is above 250 °C. The LP turbine is partly wet and the Wilson line is at less than 78 °C. It should be noted that SCC of HP turbine discs and blades has not been reported, despite condensation occurring at very high temperature. This is probably due to the fact that the loading level of HP turbine discs and blades is much lower than that of the LP turbines.
The thickness and continuity of the liquid film depend on the steam moisture content, chemical impurities, wettability and rotation speed of the surface where the film is located *°**. Film thicknesses of up to 120 xm have been measured in model turbines at MEI°. An EPRI” study revealed that at steam moisture less than 1%, no continuous liquid film forms on the turbine components in the steam flow path. However, early condensates form at moisture levels of less than 1%. The distribution of temperature and pressure in the turbines is complex and varies from station to station, as described above. Therefore, it is not possible to make general statements on the distribution of steam moisture content in the turbine and hence the exact location where condensation starts to occur. The temperatures at which the first condensation occurs in the LP, IP and HP turbines for various stations in the UK are summarised in Table 8. However, it should be noted that there are variations in the distribution of temperature and pressure in the turbines during operation, especially during start-up and shutdown period. Therefore, the Wilson line may move. Hence, there are locations that will experience cyclic dry/wet steam conditions. Furthermore, early condensate could form above the saturation temperature due to local supercooling of the steam flow and the presence of impurities”.
The thickness and continuity of the liquid film depend on the steam moisture content, chemical impurities, wettability and rotation speed of the surface where the film is located *°**. Film thicknesses of up to 120 xm have been measured in model turbines at MEI°. An EPRI” study revealed that at steam moisture less than 1%, no continuous liquid film forms on the turbine components in the steam flow path. However, early condensates form at moisture levels of less than 1%. The distribution of temperature and pressure in the turbines is complex and varies from station to station, as described above. Therefore, it is not possible to make general statements on the distribution of steam moisture content in the turbine and hence the exact location where condensation starts to occur. The temperatures at which the first condensation occurs in the LP, IP and HP turbines for various stations in the UK are summarised in Table 8. However, it should be noted that there are variations in the distribution of temperature and pressure in the turbines during operation, especially during start-up and shutdown period. Therefore, the Wilson line may move. Hence, there are locations that will experience cyclic dry/wet steam conditions. Furthermore, early condensate could form above the saturation temperature due to local supercooling of the steam flow and the presence of impurities”.
Figure 8. Mollier diagram showing regions of deposition and condensation and their effect on corrosion’. When the steam enters the turbine, the impurities from the steam can be transported onto the metal surface by deposition (vapour to solid) and condensation (vapour to liquid). The deposits can be formed on the LP turbine, as well as on the HP and IP turbines, especially during major chemical excursions. Early condensates form even when the steam wetness is less 1% due to the local supercooling and presence of impurities®. When the steam wetness is above 1%, a continuous liquid film may form on the turbine surface. The impurities formed on the surface of blades and discs reduce the turbine efficiency and cause pitting, erosion and stress corrosion cracking, as shown in the Mollier diagram’.
Figure 8. Mollier diagram showing regions of deposition and condensation and their effect on corrosion’. When the steam enters the turbine, the impurities from the steam can be transported onto the metal surface by deposition (vapour to solid) and condensation (vapour to liquid). The deposits can be formed on the LP turbine, as well as on the HP and IP turbines, especially during major chemical excursions. Early condensates form even when the steam wetness is less 1% due to the local supercooling and presence of impurities®. When the steam wetness is above 1%, a continuous liquid film may form on the turbine surface. The impurities formed on the surface of blades and discs reduce the turbine efficiency and cause pitting, erosion and stress corrosion cracking, as shown in the Mollier diagram’.
Figure 9. Chloride content in the water phase as a function of steam wetness at various cation conductivities of the inlet steam. It can be seen that the chloride concentration in the liquid phase at 1% of moisture, at which a continuous liquid film may form, is 1.6 ppm under normal operating conditions (cation conductivity of the steam equal to 0.2 uS/cm) and can be as high as 10 ppm during a chemistry excursion (cation conductivity equal to 1 uS/cm). At the Wilson line where the moisture content is 3 — 4 %, the chloride concentration in the liquid film is approximately 0.45 ppm, 1.2 ppm and 2.4 ppm when the cation conductivity of the steam is 0.2 uS/cm, 0.5 uS/cm, 1.0 uS/cm respectively. However, From Equation (8), we obtain (as y should be a positive value)
Figure 9. Chloride content in the water phase as a function of steam wetness at various cation conductivities of the inlet steam. It can be seen that the chloride concentration in the liquid phase at 1% of moisture, at which a continuous liquid film may form, is 1.6 ppm under normal operating conditions (cation conductivity of the steam equal to 0.2 uS/cm) and can be as high as 10 ppm during a chemistry excursion (cation conductivity equal to 1 uS/cm). At the Wilson line where the moisture content is 3 — 4 %, the chloride concentration in the liquid film is approximately 0.45 ppm, 1.2 ppm and 2.4 ppm when the cation conductivity of the steam is 0.2 uS/cm, 0.5 uS/cm, 1.0 uS/cm respectively. However, From Equation (8), we obtain (as y should be a positive value)
The effect of the relative exhaust steam (ratio of the exhaust steam flow to the inlet steam flow) on the partitioning of gaseous species can be illustrated by the partitioning of oxygen. Figure 12 shows the oxygen concentration in the condensate partitioned from the steam containing 18 ppm oxygen at 283 °C and 184°C as a function of the relative exhaust steam flow’ (the reason for the choice of 18 ppm oxygen was not explained). At near 100% exhaust steam flow, the concentrations of oxygen in the condensate are consistent with those calculated based on an equilibrium system. When the relative exhaust steam is close to zero, the oxygen concentrations in the condensate approaches the value in the inlet steam.
The effect of the relative exhaust steam (ratio of the exhaust steam flow to the inlet steam flow) on the partitioning of gaseous species can be illustrated by the partitioning of oxygen. Figure 12 shows the oxygen concentration in the condensate partitioned from the steam containing 18 ppm oxygen at 283 °C and 184°C as a function of the relative exhaust steam flow’ (the reason for the choice of 18 ppm oxygen was not explained). At near 100% exhaust steam flow, the concentrations of oxygen in the condensate are consistent with those calculated based on an equilibrium system. When the relative exhaust steam is close to zero, the oxygen concentrations in the condensate approaches the value in the inlet steam.
Figure 14. Oxygen concentration on the metal surface (x = 0) as a function of time at various values of condensate layer thickness.
Figure 14. Oxygen concentration on the metal surface (x = 0) as a function of time at various values of condensate layer thickness.
The calcu function o 15 and 16. contains | pH when t the liquid ro ated concentration of CO, in the liquid film and the effect on pH as a temperature at various contents of CO in the steam are shown in Figures It can be seen that the CO, in the liquid film is about 1 ppb if the steam .) ppm CO». There is no effect of the dissolved COz in the liquid film on he content of CO in the steam is 1.5 ppm. The effect of CO, on the pH of film is still negligible even if the CO, content in the steam increases to 10 ppm. The effect of such a low concentration of dissolved CO, on the conductivity of the cond ensates is also negligible.
The calcu function o 15 and 16. contains | pH when t the liquid ro ated concentration of CO, in the liquid film and the effect on pH as a temperature at various contents of CO in the steam are shown in Figures It can be seen that the CO, in the liquid film is about 1 ppb if the steam .) ppm CO». There is no effect of the dissolved COz in the liquid film on he content of CO in the steam is 1.5 ppm. The effect of CO, on the pH of film is still negligible even if the CO, content in the steam increases to 10 ppm. The effect of such a low concentration of dissolved CO, on the conductivity of the cond ensates is also negligible.
Impurities in early condensates
Impurities in early condensates
At a chloride level in the inlet steam similar to that under normal operating conditions (2 ppb to 3 ppb), the chloride concentration in the early condensate was approximately 100 ppb, which is similar to that measured at real power stations’’. It should be noted that the chloride concentration is less than that calculated based on the equilibrium data (see previous section), assuming the moisture content is 1% or less (Figure 9). This suggests that the impurities in the steam and the early condensate are not at equilibrium, i.e. there are still some impurities remaining in the steam. ‘igure 18. Chloride content in the early condensate and the liquid film as a function of chloride content in the inlet steam (EC: Early Condensate, LF: liquid film, PT: Phosphate Treatment, OT: Oxygen Treatment)”.
At a chloride level in the inlet steam similar to that under normal operating conditions (2 ppb to 3 ppb), the chloride concentration in the early condensate was approximately 100 ppb, which is similar to that measured at real power stations’’. It should be noted that the chloride concentration is less than that calculated based on the equilibrium data (see previous section), assuming the moisture content is 1% or less (Figure 9). This suggests that the impurities in the steam and the early condensate are not at equilibrium, i.e. there are still some impurities remaining in the steam. ‘igure 18. Chloride content in the early condensate and the liquid film as a function of chloride content in the inlet steam (EC: Early Condensate, LF: liquid film, PT: Phosphate Treatment, OT: Oxygen Treatment)”.
Table 9 Chemistry of early condensate and liquid film”. EC is the early condensate, LF-II and LF-III are liquid film (sections II and IID)
Table 9 Chemistry of early condensate and liquid film”. EC is the early condensate, LF-II and LF-III are liquid film (sections II and IID)
[able 10 Chemical analyses of deposits on blades failed from SCC". Steam condensation occurs on both TAI and TA2 Blade surface on the inlet steam side Blade surface on the outlet steam side
[able 10 Chemical analyses of deposits on blades failed from SCC". Steam condensation occurs on both TAI and TA2 Blade surface on the inlet steam side Blade surface on the outlet steam side
Figure 19. Deposition pattern found in Fossil Cycle turbines*’. would suggest that the real contamination of impurities would be much higher than that calculated based on steady flow and under normal operating conditions. However, it is also possible that the deposits are accumulated during major chemical excursions. In addition, a build-up of deposits is also likely in the locations near the Wilson lines since these may move due to variation in the distribution of temperature and pressure in the turbines, resulting in local dry/wet steam cycles. These deposits may form some solid products and hence do not dissolve in the condensate.
Figure 19. Deposition pattern found in Fossil Cycle turbines*’. would suggest that the real contamination of impurities would be much higher than that calculated based on steady flow and under normal operating conditions. However, it is also possible that the deposits are accumulated during major chemical excursions. In addition, a build-up of deposits is also likely in the locations near the Wilson lines since these may move due to variation in the distribution of temperature and pressure in the turbines, resulting in local dry/wet steam cycles. These deposits may form some solid products and hence do not dissolve in the condensate.
sure 1. Locations and orientations of cracks found in low pressure turbine discs in US power plants (after McMinn et al'). McMinn et al' reported that cracks on nuclear shrunk-on discs were found mainly at four locations: keyways, bores, web faces and rim attachment areas (Figure 1), with many of the failures initiated at some type of crevice (with crevice tightness a factor) and all were downstream of the location where the steam condenses in the turbine. Cracking was intergranular, along the prior austenite grain boundaries. Also, there was a Strong relationship between cracking and the concentration of oxygen in the water-steam system of power plants.
sure 1. Locations and orientations of cracks found in low pressure turbine discs in US power plants (after McMinn et al'). McMinn et al' reported that cracks on nuclear shrunk-on discs were found mainly at four locations: keyways, bores, web faces and rim attachment areas (Figure 1), with many of the failures initiated at some type of crevice (with crevice tightness a factor) and all were downstream of the location where the steam condenses in the turbine. Cracking was intergranular, along the prior austenite grain boundaries. Also, there was a Strong relationship between cracking and the concentration of oxygen in the water-steam system of power plants.
Figure 3. Crack growth rates estimated from service observations (after Speidel and Magdowski’).
Figure 3. Crack growth rates estimated from service observations (after Speidel and Magdowski’).
Figure 4. Crack growth dependence on temperature. These are claimed to be close to data from laboratory tests in aerated deionised water as shown in Figure 4.
Figure 4. Crack growth dependence on temperature. These are claimed to be close to data from laboratory tests in aerated deionised water as shown in Figure 4.
The lack of dependence of crack growth rate on the stress intensity factor, suggested by Figure 7, might seemed to be queried by work of McIntyre et al'* on 3.5 NiCrMoV steel. An increase in growth rate from 4.2x10° mm s' to 6.4x10° mm s" was reported when the K value was increased from 46 MPa m°° to 69 MPa m°°. Tests were carried out in condensing steam at 95 °C but the extent of repeatability of the data was not indicated.
The lack of dependence of crack growth rate on the stress intensity factor, suggested by Figure 7, might seemed to be queried by work of McIntyre et al'* on 3.5 NiCrMoV steel. An increase in growth rate from 4.2x10° mm s' to 6.4x10° mm s" was reported when the K value was increased from 46 MPa m°° to 69 MPa m°°. Tests were carried out in condensing steam at 95 °C but the extent of repeatability of the data was not indicated.
For high strength steels, there was no benefit from steel cleanliness. Indeed, Holdsworth et al'* reported an inferior resistance to SCC of superclean steel of 875 MPa strength level when tested at stresses close to yield in a condensing steam environment at 95 °C.
For high strength steels, there was no benefit from steel cleanliness. Indeed, Holdsworth et al'* reported an inferior resistance to SCC of superclean steel of 875 MPa strength level when tested at stresses close to yield in a condensing steam environment at 95 °C.
vessel for 8 h at 70 °C gave similar results within the scatter band established for pure condensing steam under nominally constant conditions. The use of an ammonia/hydrazine treatment gave a discernible improvement in combination with the two-shifting environment but not in condensing steam on its own. Increasing the conductivity (by addition of NazSO, into the steam line as a fine droplet spray after the boiler) or pH had no impact on crack growth in condensing steam, whilst MoS» lubricant contamination also had no effect, the rational being that in the WOL test the lubricant was confined to the notch and had no impact on crack tip chemistry. This is in contrast to the behaviour for plain specimens where a dramatic effect of MoS: is observed (Figure 14). Here also a clear effect of solution conductivity can be discerned, whilst for the same conductivity pH does not seem to have an effect. In testing by Kondo et al”? on 3.5% NiCrMoV steels, an increase in solution conductivity decreased the number of cycles to crack initiation (defined at a crack depth of 0.5 mm), Figure 15, but had no impact on the crack propagation rate (Figure 16), in this 100 ppb oxygenated solution at 60 °C, consistent with the observations of Holdsworth et al’”.
vessel for 8 h at 70 °C gave similar results within the scatter band established for pure condensing steam under nominally constant conditions. The use of an ammonia/hydrazine treatment gave a discernible improvement in combination with the two-shifting environment but not in condensing steam on its own. Increasing the conductivity (by addition of NazSO, into the steam line as a fine droplet spray after the boiler) or pH had no impact on crack growth in condensing steam, whilst MoS» lubricant contamination also had no effect, the rational being that in the WOL test the lubricant was confined to the notch and had no impact on crack tip chemistry. This is in contrast to the behaviour for plain specimens where a dramatic effect of MoS: is observed (Figure 14). Here also a clear effect of solution conductivity can be discerned, whilst for the same conductivity pH does not seem to have an effect. In testing by Kondo et al”? on 3.5% NiCrMoV steels, an increase in solution conductivity decreased the number of cycles to crack initiation (defined at a crack depth of 0.5 mm), Figure 15, but had no impact on the crack propagation rate (Figure 16), in this 100 ppb oxygenated solution at 60 °C, consistent with the observations of Holdsworth et al’”.
Figure 16. Impact of solution conductivity on crack growth kinetics (after Kondo et al”’).
Figure 16. Impact of solution conductivity on crack growth kinetics (after Kondo et al”’).
Figure 19. Effect of CO2 on stress corrosion crack growth kinetics of rotor disc steel (after Speidel et al*”*).
Figure 19. Effect of CO2 on stress corrosion crack growth kinetics of rotor disc steel (after Speidel et al*”*).
Figure 20. Crack initiation testing on LP turbine rotor and disk materials”,
Figure 20. Crack initiation testing on LP turbine rotor and disk materials”,
Figure 21. Stress corrosion crack initiation of LP turbine rotor and disc materials”.
Figure 21. Stress corrosion crack initiation of LP turbine rotor and disc materials”.
22. Impact of environmental variables on stress corrosion crack growth kinetics (after McMinn et al’).
22. Impact of environmental variables on stress corrosion crack growth kinetics (after McMinn et al’).
Surface state Clearly, the presence of crevices adjacent to the stressed surface increases the likelihood of cracking (see Discussion section).
Surface state Clearly, the presence of crevices adjacent to the stressed surface increases the likelihood of cracking (see Discussion section).
gure 28. Measured charge passed at constant potential during corrosion fatigue testing with variable hold time (after Endo et al''). Using a 0.01% NaCl environment to give reasonable conductivity (compared with the basis for Figure 27) they also measured the impact of hold time on the charge passed at the notch root (sides shielded) at maximum load as shown in Figure 28. This involved conducting tests in which the current was measured in each cycle with the cycle hold time increased progressively in sequence. The electrode potential was not given and the extent to which this varied with cycle hold time not described.
gure 28. Measured charge passed at constant potential during corrosion fatigue testing with variable hold time (after Endo et al''). Using a 0.01% NaCl environment to give reasonable conductivity (compared with the basis for Figure 27) they also measured the impact of hold time on the charge passed at the notch root (sides shielded) at maximum load as shown in Figure 28. This involved conducting tests in which the current was measured in each cycle with the cycle hold time increased progressively in sequence. The electrode potential was not given and the extent to which this varied with cycle hold time not described.
igure 29. Impact of hold time and strain rate on the number of fatigue cycles to crack initiation (after Kondo et al”).
igure 29. Impact of hold time and strain rate on the number of fatigue cycles to crack initiation (after Kondo et al”).
igure 30. Impact of hold time on corrosion fatigue crack propagation kinetics(after Kondo et al”’).
igure 30. Impact of hold time on corrosion fatigue crack propagation kinetics(after Kondo et al”’).
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved