A core part of many modern cogeneration/CHP units is the heat recovery steam generator (HRSG). However, Brad Beucker warns that the industry is ignoring flow-accelerated corrosion in HRSGs, which are particularly susceptible, and is doing so at its peril.

Single-phase FAC has a distinctive ‘orange peel’ texture Credit: D. Johnson, ChemTreat

Combined-cycle power/steam generating units are becoming increasingly popular as the energy source for many cogeneration/CHP projects.

The core of any such unit is commonly a combustion turbine, but additional power and/or process steam is generated via heat recovery steam generators (HRSGs) to feed steam turbines and process heat exchangers. Virtually all modern combined-cycle plants do not contain copper alloys, but a mindset that seems to be impossible to extinguish is the use of an oxygen scavenger/metal passivator reducing agent for condensate/feedwater treatment. This is despite the fact that for nearly three decades, this chemistry has been known to cause flow-accelerated corrosion (FAC), where in some cases such induced failures have resulted in serious injury or caused even fatalities at power plants. FAC is of particular concern in HRSGs because of their many tight-radius elbows.

Current thinking on FAC

I recently attended the spring 2013 meeting of the ASME Research Committee on Power Plant & Environmental Chemistry. One of the main topics was once again FAC and its prevention. The issue has not lost any importance since four workers were killed by an FAC-induced failure in 19861, with a number of fatalities since. In fact, FAC is the top corrosion mechanism in HRSG, so the issue has, if anything, become more pressing.

However, as was pointed out by several of the meeting attendees – and most notably by Dr. Barry Dooley of Structural Integrity Associates, and formerly of EPRI – concern over FAC seems to be fading away in the minds of plant management at many facilities. A contributing factor undoubtedly is the many retirements in the power industry, while new personnel simply do not understand the importance of FAC control. Yet, FAC continues to generate international conferences, the most recent of which was held in March in Washington, D.C., US. In a presentation to the PPEC meeting attendees, Kevin Shields, one of Dr. Dooley’s colleagues, provided the following statement in an introductory slide.2

Figure 1. Demonstrating the tube-wall thinning caused by single-pase FACCredit: D. Johnson, ChemTreat

“FAC occurs in >70% of fossil plants and represents >40% of all tube failures in HRSGs, despite R&D since the 1960s, many hundreds of plant assessments worldwide, numerous fatalities and serious failures, and much application and development.”

In this article I focus on FAC and methods to prevent it, and hopefully it will serve as a warning for plant management at the many hundreds of facilities that continue to be constructed and brought on line, not only in the US but worldwide.

When I began my utility career in 1981, conventional wisdom said that any dissolved oxygen (DO) which entered the condensate/feedwater system of utility boilers was harmful. At that time, over 50% of the power produced in the US came from coal. Coal-fired units typically have complex condensate/feedwater networks with numerous feedwater heaters. The prevalent thinking was that any trace of DO would cause corrosion, and indeed oxygen corrosion can be very problematic in uncontrolled situations. Therefore, virtually all feedwater systems for high-pressure steam generators were equipped with a deaerator for DO removal. A properly operating deaerator can lower DO concentrations to as low as 7 µg/l.

However, any residual DO concentration was still considered harmful, so chemical deaeration was a standard process at most plants. The workhorse for many years was hydrazine (N2H4), a reducing agent which reacts with oxygen as follows:

Also, arguably the primary benefit of hydrazine is that it will passivate oxidized areas of piping and tube materials as follows:

Magnetite (Fe3O4), is a protective layer that forms on carbon steel when it is placed into service. Cu2O forms on copper alloys, although we will not discuss this chemistry in great depth because the use of copper alloys in condensate/feedwater systems has greatly diminished in large part due to the potential for copper carryover to steam in high-pressure utility boilers.

Hydrazine residuals were typically maintained at relatively low levels of perhaps 20–100 µg/l (ppb). Oxygen scavenger treatment was coupled with a feed of ammonia or an amine to maintain feedwater pH within a mildly alkaline range, 9.1 to 9.3 for mixed-metallurgy feedwater systems and a bit higher for all-ferrous systems.

This programme became known as all-volatile treatment reducing [AVT(R)].

Due to the suspected carcinogenic nature of hydrazine, alternative chemicals such as carbohydrazide, methyl ethyl ketoxime, and others gained popularity. Regardless, all still had the same purpose, to establish a reducing environment in the feedwater circuit, thus inhibiting oxidation of metal. The technique became a standard in the industry.

This changed in 1986. On 9 December of that year, an elbow in the condensate system ruptured at the Surry nuclear power station, near Rushmere, Virginia, US. The failure caused four fatalities and tens of millions of dollars in repair costs and lost revenues.1

However, researchers learned from this accident and others that the reducing environment produced by oxygen scavenger feed results in single-phase FAC.

The attack occurs at flow disturbances such as elbows in feedwater piping and economizers, feedwater heater drains, locations downstream of valves and reducing fittings, attemperator piping, and, most notably for the combined-cycle industry, in low-pressure evaporators. The effect of single-phase FAC is outlined in Figure 1.

Metal loss occurs gradually until the remaining material at the affected location can no longer withstand the process pressure, whereupon catastrophic failure occurs. The thinning is due to the combination of a reducing environment and localized fluid flow disturbances, which cause dissolution of ferrous ions (Fe+2) from the metal and metal oxide matrix.

Results from EPRI show that iron dissolution is greatly influenced by not only reducing conditions but also by solution pH and temperature.

As Figure 2 illustrates, corrosion reaches a maximum at 150oC. Thus, feedwater systems and HRSG low-pressure evaporators are particularly susceptible locations. Also note the influence of pH, as reflected by ammonia concentration, on the corrosion characteristics. As we shall see, this factor is quite important with regard to control of FAC.

The quest to maintain a non-detectable oxygen residual in feedwater systems led to FAC at many coal-fired power plants. I observed this first hand at one of two utilities in which I was employed in the past. At this plant, a feedwater heater drain line failed due to FAC, shutting down an 800 MW supercritical unit. Infinitely more serious was FAC-induced failure of an attemperator line in 2007 at another of the utility’s stations, which killed two workers and seriously injured a third.

In large measure, coal plant personnel have recognized the problem of single-phase FAC, and have adopted alternative feedwater treatment methods to mitigate the issue. However, I regularly review combined-cycle proposals in which the developer specifies an oxygen scavenger feed system for HRSG chemistry control. It is obvious that this mindset clearly has not been expunged at many locations.

A flow scematic of a triple-pressure HRSG


FAC mindset change

HRSGs by their very nature have many waterwall tubes with short-radius elbows. Thus, the HRSG contains numerous spots susceptible to single-phase FAC. A primary method to mitigate this attack is the selection of proper feedwater treatment, which will be examined below.

Over 40 years ago, researchers in Germany and then Russia began using a programme known as oxygenated treatment (OT) to minimize carbon steel corrosion and iron dissolution in supercritical steam generators. The key component of the program mewas, and still is, deliberate injection of pure oxygen into the condensate/feedwater network to establish oxygen residuals of up to 300 µg/l. What chemists discovered is that in very pure feedwater (cation conductivity, ≤0.15 µS/cm), the oxygen causes the magnetite to develop a tenacious and very insoluble film of ferric oxide hydrate (FeOOH).

Results quickly showed that OT can lower feedwater iron concentrations to 1 ppb or less, and greatly minimize single-phase FAC. Now, OT is the preferred feedwater treatment for once-through utility steam generators around the world. Common in the US is an oxygen residual range of 30 ppb to 150 ppb, with a recommended pH range of 8.0 to 8.5. OT has been applied to a few drum units, where EPRI guidelines call for a feedwater pH range of 9.0 to 9.4, with a DO concentration ranging from 30 µg/l to 150 µg/l.

Although OT has been successfully applied to drum boilers, another programme has evolved that is very popular for condensate/feedwater in these steam generators. It is known as all-volatile treatment [AVT(O)]. With AVT(O), oxygen is not deliberately injected into the condensate, but rather the amount that enters from condenser air in-leakage (per normal conditions – we will examine ‘normal’ shortly) is allowed to remain without any oxygen scavenger/metal passivator treatment. It should be noted at this point that OT or AVT(O) are not permissible for feedwater systems containing copper alloys, as the oxygen would simply be too corrosive to the metal. Thus, in the following we will focus on AVT(O) for all-ferrous systems.

When researchers developed AVT(O), they took into account the pH effect on carbon steel dissolution, as illustrated in Figure 4. AVT(O) guidelines evolved to the following parameters:

  • Recommended pH range: 9.2–9.6
  • Feedwater DO concentration: 1–10 µg/l

As with OT, the condensate in an AVT(O) programme must be quite pure to allow oxygen to generate the

FeOOH protective layer rather than cause pitting. However, the cation conductivity upper limit with AVT(O) is a bit more relaxed at ≤0.2 µS/cm.

A relatively new twist has emerged regarding AVT(O) philosophy. Chemists have discovered that the heretofore established limit of 10 µg/l DO in the feedwater may allow single-phase FAC at some locations in feedwater systems where flow effects appear to prevent the dissolved oxygen from reaching the metal surface. A properly treated system will have a very pronounced reddish color, and if the treatment is not complete, areas of black magnetite will still be visible.

How to minimize FAC

Elevated pH also has a beneficial effect in mitigating FAC. Thus, the guidelines for feedwater pH now recommend a range of 9.2 to 9.6. With EPRI’s phosphate continuum programme or with caustic treatment alone, the drum pH can be controlled within a range of 9–10 quite readily. However, a complication sometimes arises because of HRSG design.

Most HRSGs are of the multi-pressure, drum, vertical tube style. In some cases, the feedwater circuit is designed such that feedwater enters each pressure circuit separately. In many others, however, the entire feedwater stream is routed to the low-pressure (LP) evaporator for heating before being distributed to the intermediate-pressure (IP) and high-pressure (HP) steam generators.

Two-phase FAC in a deaerator Credit: T. Gilchrist Tri-State G&T (retired)

For this configuration, phosphate or caustic feed to the LP circuit is not permissible due to the downstream effects on attemperator chemistry, and IP and HP economizers

In these situations, LP pH control is highly dependent upon the ammonia injected into the feedwater. If the condenser is tubed with ferrous materials, the pH may be taken higher than the 9.2 to 9.6 range listed above without ill effects. However, copper-alloy tubes would suffer corrosion at higher ammonia concentrations.

For new HRSGs, single-phase FAC control can also be addressed by materials selection. The addition of a small amount of chromium in the material at FAC-susceptible locations virtually eliminates the corrosion. One example is LP waterwall elbows. Fabrication of the elbows from 1¼ chromium alloy can provide great benefit. While the incorporation of this alloy adds some cost to the project, the materials are quite resistant to FAC.

Two-Phase FAC

Many steam generators, regardless of type, are susceptible to two-phase FAC. As the name implies, this corrosion mechanism occurs where water flashes to steam, resulting in a mixed-phase fluid.

For conventional units, feedwater heater shells and heater drains are common locations for two-phase FAC, but this equipment is not common for HRSGs. However, deaerators also experience two-phase fluid flow. As fluid flashes upon entering a deaerator, oxygen departs with the steam. Thus, the water that impinges upon metal surfaces does not maintain an oxidizing environment. Also, the pH of entrained water droplets within the steam is usually lower than the bulk water pH. The combination of these factors often initiates FAC.

As has been noted previously, elevated pH will help to mitigate FAC, but the HRSG configuration dictates the maximum treatment allowed. If the LP system is utilized for the heating of feedwater to the IP and HP circuits, solid alkali treatment (tri-sodium phosphate or caustic) of the LP circuit is not permissible. Control of pH can only be accomplished by ammonia, but it should be noted that ammonia hydrolysis, as previously outlined in Eq.4, decreases with rising temperature.

As with single-phase FAC, a method to combat two-phase FAC is fabrication of susceptible locations with chromium-containing steel. Again, however, this adds cost to the project.

Conclusion

FAC is an issue to be taken very seriously. I continue to see a large number of power plant proposals that still call for an oxygen scavenger feed system, and this is of concern.

In addition to the references in this article, I also encourage readers to access the web site of the International Association for the Properties of Water and Steam (www.IAPWS.org). This group, in which Dr. Dooley is one of the directors, offers free downloadable and cutting-edge technical information regarding power plant water/steam chemistry.

References

1. Guidelines for Controlling Flow-Accelerated Corrosion in Fossil and Combined Cycle Plants, EPRI, Palo Alto, CA, US: 2005. 1008082.

2. K. Shields, International Conference on Flow-accelerated Corrosion in Fossil, Combined-Cycle/HRSG and Renewable Energy Plants at the 2013 spring meeting of the ASME Research Committee on Power Plant & Environmental Chemistry, April 15-17, Houston, TX, US.

Brad Buecker is a process specialist with Kiewit Power Engineers Co., US. www.kiewit.com

More COSPP Articles
Past COSPP Issues