Electromagnetic Eddy Current Sensors for Evaluation of Sea-Cure and 2205 Duplex Condenser Tubing. Tube section containing impingement pits in a steam eroded condenser tube section. Also, details. Electric Power Research Institute (EPRI), 1300 W.T. Harris Blvd., Charlotte, NC 28262.
DESALINATION ELSEVIER Desalination 133 (2001) 149- 153 www.elsevier.com/locate/desal A new form of titanium condenser tube droplet erosion M.E. El-Dahshan”*, Liufu Wang”, A.M. Shams El Din”, Badr Bin Ashoorb “Material Testing Laboratory, Abu Dhabi Water and Electricity Authority, Research Center, PO Box 5411 I, Abu Dhabi, UAE Fax +971 (2) 508-1506 bCorrosion and Inspection Section, Al Taweelah Power Company, Abu Dhabi, UAE Received 27 June 2000; accepted 6 September 2000 Abstract A description is given of a new form of droplet erosion of titanium condenser tubes. At the sites of maximum impact by wet steam, the tube wall thickness is reduced to zero, and the tubes crack open under pressure of the cooling water. The crack is unbranched and progresses in a stepwise fashion. Along one side of the crack a series of small, circular metallic protrusions develops. Microscopic examination reveals the absence of any corrosion products on the surface. Similarly, no feather-like structure indicative of hydride formation is noted. It is assumed that impact with wet steam produces high-pressure transients and initiates high flow away from the points of impact. Counter-measures against droplet erosion are briefly reviewed and involve the reduction of steam velocity and/or degree of wetness. Plugging of the defected tubes is another practical means, and the use of protective sleeves might also be considered. Retubing with a more erosion-resistant material might present an ultimate, more feasible solution. Keywords: Titanium condenser tubes; Droplet erosion; Cracks; Circular protrusions; Counter-measures 1. Introduction Titanium condenser tubes exhibit outstanding resistance towards general pitting and crevice and stress corrosion cracking under a wide range of operating conditions [I]. They also withstand fresh, brackish and seawater flow velocities up to 20m.s-‘. Titanium exhibits these properties due to the presence on its surface of a thin, *Corresponding author. transparent, pore-free oxide film, which isolates the active metal from its surroundings. This film imparts a high positive potential to the metal, which guards against galvanic attack [2]. In fact, titanium tubes induce the corrosion of Cu-base tube plates, to which they are commonly mounted [3]. The tube plates must be cathodically protected either through sacrificial anodes [4] or by impressed currents [5]. Care must be taken, however, not to shift the titanium potential 001 l-9164/01/$- See front matter 0 2001 Elsevier Science B.V. All rights reserved PII: SOOll-9164(01)00093-5 150 ME. El-Dahshan et al. /Desalination towards too negative values, allowing the evolution of HZ [3]. The absorption of the gas leads to the formation of a variety of titanium hydrides [6], causing the disintegration of the metal and the loss of mechanical integrity. Similarly, both titanium and its oxide are readily attacked by dilute mineral acids [7], and the process of acid-wash of titanium-tubed condensers must be carried out conscientiously [7]. Recently, attack on Ti-condenser tubes from the steam side by entrained water droplets has been reported [&lo]. This attack assumes two distinct forms: general and localized. The general attack affects a wide area of the surface and results in an “emery-cloth” appearance. The second, more localized, form is manifested in tiny, pin-hole perforations, difficult to spot with the naked eye 133 (2001) 149-153 where V is the local drop velocity, n a constant having the value 4-6, e, the local droplet diameter, m the constant having the value 1-2, Q the number of droplets impinging per unit area and unit time, Rinc the resistance of material against initiation of droplet erosion, and &,ropis the resistance of material against propagation of droplet erosion. Theoretical calculations for a number of condensers suggested erosion to involve bombardment with droplets of mean diameter between 300-1300 pm, moving at velocities ranging between 40-120 ms-’ [S]. In the present paper we report on a new form of steam-side droplet erosion of titanium condenser tubes which, as far as we are aware, has not been previously encountered. CSILike other surface processes, droplet erosion of Ti tubes is considered to occur in two distinct stages: incubation and propagation. The incubation period may be expressed either as time or asnumber of impacts (droplets per unit area) [8] counted from the moment of exposure to the steam until the appearance of detectable defects. The incubation period is considered to be the time needed to strain harden the surface layer of the metal, to the extent that it becomes sufficiently brittle for a defect to occur. It depends on impact velocity, droplet size and material characteristics. By the end of the incubation period, a steady rate of attack, the propagation rate will be attained. The propagation rate also depends on impact velocity and droplet size. Empirical relations describing the incubation time and the erosion (propagation) rate are given as [8]: l incubation time ~1Ri,l(V”. l em. Q) (1) erosion rate u V”. am.Q. &mp (2) 2. Case description The Taweelah Power Company (Abu Dhabi, UAE) runs four large MSF distillers, each producing 12.4mgd of potable water. The vent condensing system of these units involves a main condenser receiving wet steam from the deaerator through two symmetrically located ducts. Steam then passes through three smaller units known successively as stages I, II and III. All condensers have welded Ti tubes with a 25mm outer diameter and 1 mm wall thickness. The tubes are cooled by once-through seawater. During the winter of 1999, approximately 30 months after start-up, the vent condensate of one of the units started to show increased conductivity, denoting tube leakage. Investigation showed the problem to originate in the main condenser. Removing the steam ducts of the deaerator revealed a violetcoloured circular area, encompassing 7-8 tubes of the first row of tubes. The middle two/three tubes, which suffered the most, carried longitudinal cracks some 40-50mm long (Fig. 1). The cracks were not associated with tube welds. From its position in the middle of the affected 3. Crack characteristics Fig. 1. General view of Ti tubes showing different stages of crack formation as result of droplet erosion. zone (Fig. l), it is only reasonable to conclude that failure was decided by steam impact angle, for which 90” represents the worst case [S]. The phenomenon of cracking occurred on both the cold and hot ends of the condenser, which were equally subjected to steam from the symmetrically situated ducts. The heat-affected tubes were plugged, and those with cracks were extracted for closer examination and laboratory testing. The above-described features were noted only in the main condenser. Examination carried out on stage I-III tubes revealed them to be intact, with no colouration under their steam ducts. Visual examination ofthe failed tubes showed them to carry single, unbranched cracks, some 40-50 mm in length. The crack was widest in the middle and tapered on both sides. Closer examination revealed the crack to proceed in a stepwise fashion (Fig. 2). Running mainly along one side of the crack was a series of tiny, circular protrusions (Fig. 2). Tube dimension measurements showed the outer tube diameter and the wall thickness of the damaged tubes to be identical to those of unaffected tubes, except in the region between 5 and 7 o’clock, where both dimensions were greatly diminished. At 6 o’clock the wall thickness was reduced to zero and the tube diameter was affected correspondingly. X-ray analysis revealed the tube material to correspond to alpha titanium alloy, commonly used for condenser tubes. Except for a slightly above normal (0.7%) iron content, no other abnormality was detected in material composition. Microscopic examination of the crack area and its surroundings indicated the absence of any corrosion products on the surface. Similarly, metallographic examination of tube material, far from the crack, at the crack tip and at one of the spherical protrusions (Fig. 3) showed the same metallic structure. No inclusions of corrosion products were detected, and no feather-like structure, indicative of titanium hydride formation, was observed. Fig. 2. Details of a crack showing stepwise propagation and circular protrusions. M.E. El-Lhhsl~~r~ et al. /Destrliti~ition 133 (20011 149-153 with a special angle of impact of the droplets. It is to be noted, however, that the crack never crosses through any of these protrusions. Apparently their dense nature counteracts crack approach and changes its path to the stepwise fashion noted. 4. Mitigation of droplet erosion of Ti tubes Fig. 3. Cross section through a spherical protrusion, revealing its metallic nature and the absence ofcorrosion products or feather-like structure. Based on the strength of the above observations and evidence, we conclude that the cracking of the Ti tubes of the vent condenser is purely mechanical in nature. The absence of any corrosion products and the confinement of the cracks to the region facing the incoming steam supportthis conclusion. For steam to produce the observed damage, it must carry tiny droplets of water (wet steam). As these emerge from the vacuum deaerator, designed to remove gases, the presence of entrapped gas bubbles in the steam is highly unlikely. The high-velocity impact of water droplets against tube surface is expected to produce high-pressure transients at the points of impact and cause high flow away from the points of impact. Both effects are damaging to the metal and are expected to lead to its destruction as tiny particulates. The small round protrusions developing along the side of the crack are of interest. As is seen in Fig. 3, these are metallic in nature; hence we conclude that they represent part of the material displaced away from the points of impact. The precise mechanism whereby this process occurs is, however, not clear. Equally obscure is why the majority of the protrusions develops on one side of the crack. It is possible that this is connected Droplet erosion of Ti tubes results from impingement with wet steam at high flow rates. Its effect is limited to the first, or maximum the second tube rows of the condenser. A reduction of steam velocity might lower or eliminate erosion. This solution may, however, be difficult to realize in already existing plants. Another approach to limit droplet erosion is to decrease the “wetness” content of the steam. This can be achieved either by regulating the temperature or through the installation ofprotective flow devices in the form of deflecting grids or shielding artifices. These methods eliminate the action of large-size, high-velocity water droplets, while allowing steam to pass through with minimum pressure drop. Apparently the passage of the wet steam through the main vent condenser stripped it of most of its entrained droplets so that the almost dry steam had no effect on the tubes of stages I-III. The most practical way of dealing with the problem of droplet erosion is the plugging of the failed tubes. Since only the first tube row is affected, their plugging will not result in the loss of a large portion of the cooling capacity of the condenser. In the case covered by the present paper, where steam is delivered to the condenser through limited ducts, the number of affected tubes will be even smaller than the full outer row. Plugging of titanium tubes should be done with neutral materials, e.g., wood, plastic or rubber. Copper-base alloys or stainless steel plugs should not be used, as they will be gaivanically attacked. M.E. El-Dahshan et al. /Desalination Protective sleeves have been extensively used in Europe during the last decade. Stainless steels are used to protect tube sections against droplet erosion, and although their use has been proven to be effective, they do cause a lowering of steam side pressure. This reduction is, however, less than that resulting from grid panels and dummy tubes. There is also a small reduction of the heat transfer at the sleeved section of the tubes. Because of its nature, droplet erosion is confined to the first, or at most the second, tube row. Should the plugging of these tubes represent a measurable loss of cooling efficiency, retubing might be considered, and in this event, tubes of a material with higher droplet-erosion resistance must be selected. A high (6.0%) MO stainless steel appears to be a suitable material for this purpose. 5. Conclusions 1. Some of the Ti tubes of the vent condensers of the MSF distillers at Taweelah Power Company (Abu Dhabi, UAE) suffered unusual attack from the steam side. The attack took the form of cracks initiated by droplet erosion. No corrosion products were detected and no evidence of titanium hydride formation was noted. 133 (2001) 149-153 153 2. The attack was most severe on tubes running perpendicular to steam flow, decreasing on either side. 3. Along the crack length, small metallic protrusions formed on one side of the crack. These are thought to result from plastic flow of the metal. 4. Ways of mitigation of droplet erosion were outlined and briefly discussed. References PI A.M. Shams El-Din, Industrial Corrosion and Corrosion Technology, KISR (Kuwait), 1996, p. 49. PI M.G. Fontana and N.D. Greene, Corrosion Engineering, McGraw-Hill, New York, 1978, p. 32. 131T. Fukuzuka, K. Shimogori, H. Satoh and F. Kamikubo, Desalination, 3 1 (1979) 389. 141T. Moroishi and H. Miyuki, Titanium 80, Vol. 4,4th Intern. Conf. On Titanium, Kyoto, 1980, p. 2713. 151J.I. Lee, P. Chung and C. H. Tsai, Corrosion 86, NACE, 1986,259. 161A.M. Shams El Din, T.M.H. Saber and A.M. Taj El Din, Desalination, 107 (1996) 265. 171A.M. Shams El Din, H.A. El Shayeb and F.M. Abd El Wahab, J. Elecroanalyt. Chem., 214 (1986) 567. J. Tavast, Proc., ASME Joint Power Generation PI Conference, Houston, Texas, 1996, p. 4. PI J. Tavast, EPRI Condenser Technology Symposium, RI, USA, 1987. [lOI N. Iwai, T. Ozeki, K. Ohi, M. Amemiya and H. Itoh, ASME, 19 (1992) 37.