By Brad Buecker, Contributing Editor
A single waterwall tube failure in a steam generator often will cost a utility at least six figures in lost production, labor and startup costs. If the failure occurs during the hot summer months, the penalty can be even greater. And if the difficulties trigger a chemical cleaning or multiple tube replacements, watch out!
Typically, the single most common failure mechanism of waterwall tubes is under-deposit corrosion. Significant strides have been made in reducing under-deposit corrosion incidents, but not all issues have been resolved, particularly with regard to iron deposition following major maintenance outages.
Iron Deposition on Waterwall Tubes
During normal steam generator operation, condensate/feedwater piping and boiler tubes develop a layer of iron oxide, which, while being a corrosion product, protects the underlying base metal against further corrosion. This protective layer may be very tight, especially where all-volatile oxygenated treatment [AVT(O)] or especially oxygenated treatment (OT) is utilized. During periods of chemistry upsets, thermal transients, and forced outages, additional corrosion products are generated. And, from the major work often performed during scheduled maintenance outages, literally hundreds to thousands of pounds of loose particulates, primarily composed of iron oxides, may collect in the condenser hotwell, condensate and feedwater systems. As these particulates enter the boiler, they precipitate on the tubes. Because the transported deposits are porous, they establish wick boiling. In this mechanism, the bulk boiler water enters the deposits and migrates toward the tube surface where the water flashes to steam while any contaminants remain behind.
|FIGURE 1 Hydrogen damage failure of a waterwall tube. Note the thick-lipped fracture with little noticeable metal loss.|
|Photo courtesy of ChemTreat.|
Consider what can happen during times of a condenser tube leak, even if the leak is considered small. Several under-deposit reaction mechanisms are possible, but very common is the reaction shown below.
MgCl2 + 2H2O + heat → Mg(OH)2↓ + 2HCl
Magnesium salts react with water to produce a magnesium hydroxide precipitate, plus hydrochloric acid. While HCl may cause general corrosion in and of itself, the compound will concentrate under deposits where the reaction of the acid with iron generates hydrogen, which in turn can lead to hydrogen damage of the tubes. In this mechanism, hydrogen gas molecules (H2), which are very small, penetrate into the metal wall where they then react with carbon atoms in the steel to generate methane (CH4),
2H2 + Fe3C → 3Fe + CH4↑
Formation of the gaseous methane and hydrogen molecules causes cracking in the steel, greatly weakening its strength. Hydrogen damage is troublesome because it cannot be easily detected. After hydrogen damage has occurred, the plant staff may replace tubes only to find that other tubes continue to rupture.
Hydrogen damage failures may occur in a matter of days, without any appreciable metal loss, following a significant condenser tube leak. The Electric Power Research Institute (EPRI) recommends immediate boiler shutdown if the pH drops to 8.0. The following case history illustrates the rapidity with which contaminant in-leakage can affect boiler water chemistry.
This first example comes from personal experience. An 80 MW unit supplied by a 1,250 psig coal-fired, cyclone boiler had just been returned to service from a scheduled autumn outage. Laboratory personnel discovered a condenser leak was allowing contaminants to enter the system, such that condensate total-dissolved-solids (TDS) concentrations at times reached 0.75 parts per million (ppm). Although lab staff requested that the boiler be taken off line immediately, operations managers declined due to load demand issues.
The boiler was on congruent phosphate control (this was in the early 1980s), so the lab staff increased monitoring frequency and attempted to maintain phosphate and pH levels within recommended guidelines. After approximately three weeks, operators discovered the source of the leak and corrected the problem. Two months later, boiler waterwall tubes began to fail with alarming frequency. The unit came off numerous times for tube repairs and in at least one instance had only been back on-line for a few hours when another tube failed. Failures happened so regularly that plant management scheduled an emergency tube replacement during the upcoming spring outage. The repairs cost over $2,000,000. The mechanism attributed to these failures was under-deposit corrosion and hydrogen damage caused by excessive sludge and scale formation.
Interestingly, the leak was not from a failed condenser tube. The condenser hotwell is equipped with a drain line that discharges to the cooling water outlet tunnel. During the autumn outage, an operator opened the line to drain the hotwell but then forgot to close the isolation valve before startup. Once the unit went on-line, the strong condenser vacuum pulled cooling water into the hotwell. Stopping the leak was easy; repairing the waterwall tubes was not.
This example is a drastic example of under-deposit corrosion, but serious difficulties may still arise although the deposits are “simply” iron oxides. Impurities from even small condenser leaks may concentrate due to wick boiling and generate acidic conditions in localized areas. A common perception is that deposition is most prevalent on the hot side of waterwall tubes, and indeed in general this is true. However, the following example shows the unpredictability of where fouling might occur.
Because deposit-forming materials typically accumulate more rapidly and in greater depth on the hot side of tubes and in areas where the flow acquires some horizontal tangent such as nose tubes, common sampling points for deposit density analysis are the areas 20 to 30 feet above the top burner elevation and/or the nose tubes themselves. However, with steam generation always expect the unexpected. In a conversation with a power plant colleague in 2009, I learned that two boilers at a plant had experienced failures of lower waterwall tubes. Tube analyses indicated overheating due to localized iron oxide buildups generated during oil firing at startups following major outages. Certainly, the total heat load in the boiler with oil guns in service is much lower than at normal, coal-fired operation, but the oil fires generated hot spots that caused substantial iron oxide deposition; a phenomenon exacerbated by the high concentrations of particulates following outages. Condensate/feedwater chemistry is AVT(O), which, during normal operation, maintains feedwater iron concentrations below 2 parts per billion (ppb). Thus, the iron deposition has been an outage/startup problem in this case.
Iron Oxide Deposit Control
At many plants the transition in condensate/feedwater chemistry from all-volatile-reducing treatment [AVT(R)], which utilizes a reducing agent, to AVT(O) or OT has greatly reduced iron corrosion and transport during normal operation. It is not uncommon to see economizer inlet iron concentrations of less than 1 ppb. But what about iron oxide control during startups?
An equipment investment that can pay for itself several times over with just the first use is a condensate particulate filter. These straightforward mechanical devices can be easily equipped with filter cartridges that remove particulates down to 1 micron in size.
The common location for a particulate filter is just after the condensate pumps, with the filter placed in a valved, bypass loop around the main condensate feed line. The device need not be full flow, as at start-up the condensate circulation is often restricted to half the full-load flow rate or perhaps even less. A filter will remove iron oxide particulates and other “crud” within a short period of time, allowing for potentially significant reductions in hold periods, as illustrated below.
At another utility, we once started up a supercritical unit following a boiler chemical cleaning. Following the standard rinses, the only method to remove remaining iron oxide and other particulates from the condensate was filtration through the deep-bed condensate polishers. Not only did this process significantly foul the polisher resin, but four days of filtration were required to reduce the suspended solids, whose original concentration was greater than 1 ppm, to the relatively low ppb concentration necessary to fire the boiler.
To alleviate this difficulty for future startups after maintenance outages and chemical cleanings, we ordered a condensate particulate filter designed to handle half of the full-load flow. The vessel, equipped with 6-micron (absolute) filter elements, was placed ahead of the condensate polishers. The filter was first utilized in 2008 at startup following another chemical cleaning, where again the initial iron oxide concentration in the condensate was above 1 ppm. Although two filter element change-outs were required to clean the condensate, the cleanup time was reduced from four days to one. The three-day startup savings paid for the filter vessel and elements several times over after just the first use.
It has become well known among power plant chemists that the traditional phosphate treatment programs of the 1960s through 1980s for drum units have fallen out of favor due to the tendency of sodium phosphates to directly precipitate on waterwall tubes. The phenomenon is commonly known as “hideout.” Not only does hideout cause difficulties with bulk boiler water chemistry and its control, but the old congruent and coordinated programs allow precipitation of acidic phosphates that directly corrode the tube metal. Furthermore, phosphate carryover into main and reheat steam caused problems, not least of which was overheating in U-bends due to phosphate collection. At most plants, and especially those with steam generators operating at 2,000 psi or greater, the favored programs now are EPRI’s phosphate continuum (PC), caustic treatment (CT), or in some cases AVT or OT. Space limitations prevent a comprehensive discussion of these alternatives, but in each case only a small amount of alkalinity is available to counteract the acidic conditions generated by even a small condenser tube leak.
While these programs have minimized or eliminated hideout issues, their use further emphasizes the need for control of iron oxide deposition. Of course, a condenser leak could quickly generate acidic conditions, but alkaline corrosion is also a possibility. Phosphate continuum and caustic treatment operate with a small amount of free sodium hydroxide, where guidelines recommend a maximum concentration of 1 ppm free caustic with PC and 1.5 to 2 ppm with CT. When waterwall tubes are clean, the small concentration of caustic is not problematic. However, an accumulation of porous iron oxide deposits can lead to under-deposit caustic gouging, where the caustic attacks both the protective iron oxide layer and the base metal underneath.
Fe3O4 + 4NaOH → 2NaFeO2 + N2FeO2 + 2H2O
Fe + 2NaOH → Na2FeO2 + H2↑
This phenomenon can be particularly problematic if a chemistry change is made without first chemically cleaning the boiler.
Control of deposition in steam generators can pay big dividends in reliability and availability of the unit. Deposit prevention often requires a combination of good chemistry control and modern equipment utilization.
Author: Brad Buecker is a contributing editor for Power Engineering.
Reference: 1. B. Buecker, “Condenser Chemistry and Performance Monitoring: A Critical Necessity for Reliable Steam Plant Operation”; Proceedings of the 60th Annual Meeting, International Water Conference, October 18-20, 1999, Pittsburgh, Penn.