Erosion corrosion in a preheater at the support plate.
Figure 2. The erosion corrosion test loop at KEMA
Figure 3 Case 2. The spindle, eroded corroded at the top
Figure 4 Case 3. Switch lever, severely attacked after 10000 hours
Figure 5 Case 5. The attacked buffle plate in the KEMA experimental boiler
Figure 6 Case 6. The corroded and non-corroded plates in the wet steam pipe
Figure 7 Case 8. High pressure water separator with "tiger skin" pattern
Figure 8 Case 10. Welded plates of PWR waterseparator with two different heats of carbon steel. One was corroded and one did not show any damage. The lower photograph is taken at a cleaned surface.
Figure 9 Case 13. Erosion corrosion in a preheater at the support plate. Wet steam has a high velocity in the support plate holes.
Figure 10 Case 14. With a high Cr equivalent erosion corrosion even behind high butt welds did not occur
Figure 11 Case 16. Steam blanketing on the inner and erosion corrosion on the outer radius
|
|
Erosion corrosion in Heat Exchangers, the Value of Material Specification Huijbregts W.M.M., Uilhoorn F., Wels H.C. (paper 47) Euromat 1997. Materials, functionality & design, Conference Maastricht, 1997 1. Introduction An inventory of corrosion failures in the period 1973 up to 1995 was made. It appeared in the Dutch electricity power industry that erosion corrosion is still an important failure cause in water-steam systems. In this period some interesting papers on the erosion corrosion research were published (1981, 1982 and 1984). Despite the high failure risk carbon-steels are not yet specified on erosion-corrosion resistance. Return On Investment calculations show that the introduction of modified steel can result in large money savings. The costs for forced unavailability of a power plant due to corrosion failures are high. So there is sufficient reason to give the highest attention on solving the erosion-corrosion problem Corrosion failures are a result of a combination of water conditioning, design parameters and material. Kastner published mathematical correlation's, based on laboratory research and practical experience, with which the erosion-corrosion rate can be, calculated (Kastner 1986). KEMA used these published correlation's too for creating a simple but very usefull computer model, called KASEC (KEMA Advice System for Erosion Corrosion). In the mathematical correlation of Kastner the Cr and Mo content of the steel (carbon steel, 15Mo3, 13CrMo44) have to be given. According to the correlation of Kastner small amounts of alloying elements (like present in carbon steels) do not influence the erosion-corrosion rate very much. However, KEMA showed that selecting carbon steels with some of the alloying elements Cu, Cr, Mo and C can control the erosion-corrosion in ordinary water steam systems. 2. Influence of alloying elementsThe erosion-corrosion resistances of 70 heats of carbon steels were determined in a jet stream of a water-steam mixture at a steam velocity of 960 m/s and oxygen content less than 2 ppb. See Figure 2. The first results were published in 1981 (Huijbregts 1981). After an exposure time of 100 hours the weight losses of the specimens were determined. These weight losses were correlated with the steel composition, which resulted in the regression formula, mentioned below. GKEMA = 90 - 160*Cu - 115*Cr - 40*Mo + 35*C The calculated GKEMA values were in the range of 40 up to 100.
Erosion-corrosion failures occur in various components in the water-steam system such as appendages, wet steam lines and water separators, low and high-pressure preheaters, economisers and evaporators. Some of these cases have been discussed already in earlier papers (Huijbregts et all 1982, 1984). After finishing the experiments in 1982, several erosion- corrosion cases were offered for further research. Because good fitting correlation equations had been established, the new samples were not tested in the erosion-corrosion loop. They were only analysed chemically as to Cr, Cu, and Mo content, and the erosion-corrosion resistances were calculated. The failures occurred in the steam-water circuit of power stations under various conditions. Hydraulic conditions and chemical water treatment can differ very much for these components. However, our experience is that in all observed erosion-corrosion failures in water-steam systems up to now the corroded steels had a GKEMA value of more than 80. With this specification a minimum Cr-equivalent has been introduced. Cr-equivalent = Cr +1.4 Cu + 0.3 Mo - 0.3 C > 0.09 To obtain an erosion-corrosion resistant carbon steel (under not too excessive erosion-corrosion conditions) the Cr-equivalent should be more than 0.09. Because small amounts of Cu, Cr and Mo improve the resistance to erosion- corrosion very much, a large variation in erosion-corrosion resistance of carbon-steel exists. From one melt of steel various products (tubes of different diameter, plates etc) with possible various heat treatments are made. Such a melt of steel is called a heat. Different heats will accord to a specification of a steel type, pa15Mo3. Accidentally one heat can possess a much higher resistance compared to other heats from the same steel specification. Thus, erosion-corrosion resistance of the material is nowadays still a matter of "Russian roulette". Erosion-corrosion was observed on steels having a low Cr-equivalent: Special attention was paid to the cases in which certain parts of the component were corroded but other parts were not damaged, even though they had been operated under the same conditions. 4. Various appendages Case 1 A number of steam sieves suffered erosion- corrosion. Chemical analyses showed that three different steel heats were applied. The calculated Cr-equivalents were 0.077, 0.080 and 0.015. The last heat was corroded the most. Case 2 A spindle from a condensate effluent valve of a low pressure preheater was coated only partly with a chromium steel (9% Cr). see Figure 3. It might be considered that on the top of the spindle the coating was not necessary because in general there are no corrosion failures. But because of the too low Cr-equivalent the carbon steel was corroded. The Cr-equivalent was 0.060. Case 3 The switch lever of a supply water pump had been attacked severely by erosion-corrosion. See Figure 4. The Cr-equivalent was -0.002. The operation time of this lever was 10,000 hours. Another lever from an identical pump at the same station showed less corrosion after 40,000 hours. The Cr, Cu, and Mo contents of this steel are slightly higher, resulting in a Cr-equivalent of 0.056. Case 4 There was severe erosion-corrosion in a control valve housing, resulting in a leakage. This steel had a high weight loss in the erosion-corrosion test, 107-mg. The Cr-equivalent was -0.016. Case 5 In the KEMA experimental boiler, erosion-corrosion occurred in a steam pressure drop vessel in which steam expanded and where the water steam jet hitted a buffle plate in the vessel. See Figure 5. After 10 years of operation, there was a leakage in the buffle plate and the vessel wall. The steel plate (10 mm thick) and the vessel wall (5 mm thick) were corroded and a leakage occurred. The Cr-equivalent of the steel plate was 0.138. For such situations (a buffle plate for protecting the vessel to the wet steam flow) steel with a higher Cr-equivalent is required. 5. Wet steam pipes and water separatorsErosion-corrosion had occurred in wet steam pipes and in the low and high-pressure water separators of a BWR. These components were made of a simple carbon steel. See Figure 6 and 7. Cases 6, 7, and 8 The low and high pressure water separators, the pre-waterseparator and the wet steam tubes in the BWR were replaced by stainless steel components in the last few years, before the erosion-corrosion tests were finished. Case 6 A pipe behind the BWR water separator was constructed by welding bent steel plates. One of those plates had been attacked severely. The other plates, both upstream and downstream, had not been attacked. The corroded plate showed a different chemical composition. The Cr-equivalent of the corroded steel was -0.013, whereas the values of the uncorroded plates were much higher, 0.156. Case 7 Other, smaller diameter wet steam tubes also showed erosion-corrosion, and the Cr equivalents of the two different heats amounted to 0.017 and 0.003. Case 8 The high-pressure water separator of the BWR also had severe erosion-corrosion. see Figure 7. Low Cr, Cu, and Mo contents were measured in the steel, resulting in a Cr-equivalent of 0.021. Besides, there is a remarkable phenomenon in this failure. The so called "tiger skin" pattern of erosion-corrosion on the inside of the pipe was also present on the outside. This phenomenon can be explained in terms of electrochemical polarisation of the corrosion reaction by diffusion of hydrogen through the steel wall. There where the millscale on the outside of the wall was not present anymore erosion-corrosion was started on the inside. Case 9 The low pressure water separator of a BWR showed erosion corrosion. Two different steel heats were used, both with low Cr-equivalents: 0 and -0.028. Case 10 One of the most obvious cases is the erosion-corrosion of a plate in a water separator of a PWR. See Figure 8. Two plates were welded together. Only one plate was corroded, whereas on the other non-corroded plate the corrosion products had deposited. The horse- shoe pattern of erosion-corrosion is very obvious. The corroded plate and the non- corroded plate had Cr-equivalents of respectively -0.012 and 0.275. 6. PreheatersCase 11 Two tubes of a backpressure unit were received; one showed severe erosion-corrosion, and the other had only some corrosion spots. The erosion-corrosion Cr-equivalent for the uncorroded tube was much higher than that of the corroded tube (0.253 and 0.037). Case 12 The preheater of a PWR had to be renewed because of severe erosion-corrosion. The preheater was dismantled row by row and 24 tubes were analysed The 24 tubes had been made from 2 different heats (Cr-equivalents of 0.039 and 0.096). It could be seen that in one row some tubes were corroded more severely than other tubes adjacent to them. The most severely attacked tubes were found to have the lower Cr-equivalent. Case 13 The preheater of a power station suffered local corrosion near the support plates. See Figure 9. The Cr-equivalent was -0.005. Coarse magnetite crystals were found adjacent to the corroded area. Precipitation of such coarse iron oxide crystals is often observed in case of erosion-corrosion. A second tube with a higher Cr-equivalent (0.162) did not show any corrosion. Case 14 In a high pressure preheater, erosion-corrosion was found downstream several butt welds. See Figure 10. The inner diameter of the 15Mo3 steels was 20 mm, and the weld heights for the three tubes were 1.8, 1.6, 1.7 and 2.0 mm. Two tubes, both downstream and with the lowest Cr-equivalents (0.137 and 0.154), were corroded severely, one having a leakage. Two tubes downstream with high Cr-equivalents (0.268 and 0.310) had no erosion-corrosion under these severe conditions. The conditions downstream such high butt welds are very unfavourable, so even the steels with high Cr-equivalents were not resistant enough. 7. EconomisersCase 15 Slight erosion-corrosion was observed on thin tubes in an economiser of a waste heat boiler. The Cr-equivalent was very low, -0.005. Case 16 There was severe erosion-corrosion in the economiser tube bends of a waste-heat boiler. See Figure 11. Steaming occurred in the economiser. This could also be concluded from some failed tubes where erosion-corrosion occurred on the outer radius and steam blanketing on the inner radius. The water film had come loose from the inner radius surface where coarse magnetite crystals had formed. Analysis of the steel shows the erosion-corrosion resistance of this steel to be low, because of the low Cr-equivalent (-0.002). 8. Conclusions
9. References
|
||
|
|
||||
